279 files changed, 8738 insertions, 5133 deletions
diff --git a/net/802/fc.c b/net/802/fc.c
index 34cf1ee014b8..1e49f2d4ea96 100644
--- a/net/802/fc.c
+++ b/net/802/fc.c
@@ -70,7 +70,7 @@ static int fc_header(struct sk_buff *skb, struct net_device *dev,
        if(daddr)
        {
                memcpy(fch->daddr,daddr,dev->addr_len);
-                return(hdr_len);
+                return hdr_len;
        }
        return -hdr_len;
 }
diff --git a/net/802/fddi.c b/net/802/fddi.c
index 3ef0ab0a543a..94b3ad08f39a 100644
--- a/net/802/fddi.c
+++ b/net/802/fddi.c
@@ -82,10 +82,10 @@ static int fddi_header(struct sk_buff *skb, struct net_device *dev,
        if (daddr != NULL)
        {
                memcpy(fddi->daddr, daddr, dev->addr_len);
-                return(hl);
+                return hl;
        }
-        return(-hl);
+        return -hl;
 }
@@ -108,7 +108,7 @@ static int fddi_rebuild_header(struct sk_buff	*skb)
        {
                printk("%s: Don't know how to resolve type %04X addresses.\n",
                       skb->dev->name, ntohs(fddi->hdr.llc_snap.ethertype));
-                return(0);
+                return 0;
        }
 }
@@ -162,7 +162,7 @@ __be16 fddi_type_trans(struct sk_buff *skb, struct net_device *dev)
        /* Assume 802.2 SNAP frames, for now */
-        return(type);
+        return type;
 }
 EXPORT_SYMBOL(fddi_type_trans);
@@ -170,9 +170,9 @@ EXPORT_SYMBOL(fddi_type_trans);
 int fddi_change_mtu(struct net_device *dev, int new_mtu)
 {
        if ((new_mtu < FDDI_K_SNAP_HLEN) || (new_mtu > FDDI_K_SNAP_DLEN))
-                return(-EINVAL);
+                return -EINVAL;
        dev->mtu = new_mtu;
-        return(0);
+        return 0;
 }
 EXPORT_SYMBOL(fddi_change_mtu);
diff --git a/net/802/hippi.c b/net/802/hippi.c
index cd3e8e929529..91aca8780fd0 100644
--- a/net/802/hippi.c
+++ b/net/802/hippi.c
@@ -152,7 +152,7 @@ int hippi_change_mtu(struct net_device *dev, int new_mtu)
        if ((new_mtu < 68) || (new_mtu > 65280))
                return -EINVAL;
        dev->mtu = new_mtu;
-        return(0);
+        return 0;
 }
 EXPORT_SYMBOL(hippi_change_mtu);
diff --git a/net/802/tr.c b/net/802/tr.c
index 1c6e596074df..5e20cf8a074b 100644
--- a/net/802/tr.c
+++ b/net/802/tr.c
@@ -145,7 +145,7 @@ static int tr_header(struct sk_buff *skb, struct net_device *dev,
        {
                memcpy(trh->daddr,daddr,dev->addr_len);
                tr_source_route(skb, trh, dev);
-                return(hdr_len);
+                return hdr_len;
        }
        return -hdr_len;
diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c
index a2ad15250575..25c21332e9c3 100644
--- a/net/8021q/vlan.c
+++ b/net/8021q/vlan.c
@@ -321,7 +321,7 @@ static int register_vlan_device(struct net_device *real_dev, u16 vlan_id)
        if (new_dev == NULL)
                return -ENOBUFS;
-        new_dev->real_num_tx_queues = real_dev->real_num_tx_queues;
+        netif_copy_real_num_queues(new_dev, real_dev);
        dev_net_set(new_dev, net);
        /* need 4 bytes for extra VLAN header info,
         * hope the underlying device can handle it.
@@ -525,6 +525,10 @@ static int vlan_device_event(struct notifier_block *unused, unsigned long event,
                break;
        case NETDEV_UNREGISTER:
+                /* twiddle thumbs on netns device moves */
+                if (dev->reg_state != NETREG_UNREGISTERING)
+                        break;
                /* Delete all VLANs for this dev. */
                grp->killall = 1;
diff --git a/net/8021q/vlan.h b/net/8021q/vlan.h
index 8d9503ad01da..b26ce343072c 100644
--- a/net/8021q/vlan.h
+++ b/net/8021q/vlan.h
@@ -25,6 +25,7 @@ struct vlan_priority_tci_mapping {
 *      @rx_multicast: number of received multicast packets
 *      @syncp: synchronization point for 64bit counters
 *      @rx_errors: number of errors
+ *      @rx_dropped: number of dropped packets
 */
 struct vlan_rx_stats {
        u64                     rx_packets;
@@ -32,6 +33,7 @@ struct vlan_rx_stats {
        u64                     rx_multicast;
        struct u64_stats_sync   syncp;
        unsigned long           rx_errors;
+        unsigned long           rx_dropped;
 };
 /**
diff --git a/net/8021q/vlan_core.c b/net/8021q/vlan_core.c
index 0eb96f7e44be..b6d55a9304f2 100644
--- a/net/8021q/vlan_core.c
+++ b/net/8021q/vlan_core.c
@@ -30,7 +30,7 @@ int __vlan_hwaccel_rx(struct sk_buff *skb, struct vlan_group *grp,
                skb->pkt_type = PACKET_OTHERHOST;
        }
-        return (polling ? netif_receive_skb(skb) : netif_rx(skb));
+        return polling ? netif_receive_skb(skb) : netif_rx(skb);
 drop:
        dev_kfree_skb_any(skb);
@@ -38,12 +38,12 @@ drop:
 }
 EXPORT_SYMBOL(__vlan_hwaccel_rx);
-int vlan_hwaccel_do_receive(struct sk_buff *skb)
+void vlan_hwaccel_do_receive(struct sk_buff *skb)
 {
        struct net_device *dev = skb->dev;
        struct vlan_rx_stats     *rx_stats;
-        skb->dev = vlan_dev_info(dev)->real_dev;
+        skb->dev = vlan_dev_real_dev(dev);
        netif_nit_deliver(skb);
        skb->dev = dev;
@@ -72,7 +72,6 @@ int vlan_hwaccel_do_receive(struct sk_buff *skb)
                break;
        }
        u64_stats_update_end(&rx_stats->syncp);
-        return 0;
 }
 struct net_device *vlan_dev_real_dev(const struct net_device *dev)
@@ -112,9 +111,12 @@ vlan_gro_common(struct napi_struct *napi, struct vlan_group *grp,
        }
        for (p = napi->gro_list; p; p = p->next) {
-                NAPI_GRO_CB(p)->same_flow =
+                unsigned long diffs;
-                        p->dev == skb->dev && !compare_ether_header(
-                                skb_mac_header(p), skb_gro_mac_header(skb));
+                diffs = (unsigned long)p->dev ^ (unsigned long)skb->dev;
+                diffs |= compare_ether_header(skb_mac_header(p),
+                                              skb_gro_mac_header(skb));
+                NAPI_GRO_CB(p)->same_flow = !diffs;
                NAPI_GRO_CB(p)->flush = 0;
        }
diff --git a/net/8021q/vlan_dev.c b/net/8021q/vlan_dev.c
index 3bccdd12a264..f6fbcc0f1af9 100644
--- a/net/8021q/vlan_dev.c
+++ b/net/8021q/vlan_dev.c
@@ -177,8 +177,8 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev,
        } else {
                skb->dev = vlan_dev;
-                rx_stats = per_cpu_ptr(vlan_dev_info(skb->dev)->vlan_rx_stats,
+                rx_stats = this_cpu_ptr(vlan_dev_info(skb->dev)->vlan_rx_stats);
-                                        smp_processor_id());
                u64_stats_update_begin(&rx_stats->syncp);
                rx_stats->rx_packets++;
                rx_stats->rx_bytes += skb->len;
@@ -225,7 +225,10 @@ int vlan_skb_recv(struct sk_buff *skb, struct net_device *dev,
                }
        }
-        netif_rx(skb);
+        if (unlikely(netif_rx(skb) == NET_RX_DROP)) {
+                if (rx_stats)
+                        rx_stats->rx_dropped++;
+        }
        rcu_read_unlock();
        return NET_RX_SUCCESS;
@@ -843,13 +846,15 @@ static struct rtnl_link_stats64 *vlan_dev_get_stats64(struct net_device *dev, st
                        accum.rx_packets += rxpackets;
                        accum.rx_bytes   += rxbytes;
                        accum.rx_multicast += rxmulticast;
-                        /* rx_errors is an ulong, not protected by syncp */
+                        /* rx_errors, rx_dropped are ulong, not protected by syncp */
                        accum.rx_errors  += p->rx_errors;
+                        accum.rx_dropped += p->rx_dropped;
                }
                stats->rx_packets = accum.rx_packets;
                stats->rx_bytes   = accum.rx_bytes;
                stats->rx_errors  = accum.rx_errors;
                stats->multicast  = accum.rx_multicast;
+                stats->rx_dropped = accum.rx_dropped;
        }
        return stats;
 }
diff --git a/net/9p/client.c b/net/9p/client.c
index dc6f2f26d023..f34b9f510818 100644
--- a/net/9p/client.c
+++ b/net/9p/client.c
@@ -61,13 +61,13 @@ static const match_table_t tokens = {
 inline int p9_is_proto_dotl(struct p9_client *clnt)
 {
-        return (clnt->proto_version == p9_proto_2000L);
+        return clnt->proto_version == p9_proto_2000L;
 }
 EXPORT_SYMBOL(p9_is_proto_dotl);
 inline int p9_is_proto_dotu(struct p9_client *clnt)
 {
-        return (clnt->proto_version == p9_proto_2000u);
+        return clnt->proto_version == p9_proto_2000u;
 }
 EXPORT_SYMBOL(p9_is_proto_dotu);
diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c
index c85109d809ca..078eb162d9bf 100644
--- a/net/9p/trans_fd.c
+++ b/net/9p/trans_fd.c
@@ -222,7 +222,7 @@ static void p9_conn_cancel(struct p9_conn *m, int err)
        }
 }
-static unsigned int
+static int
 p9_fd_poll(struct p9_client *client, struct poll_table_struct *pt)
 {
        int ret, n;
diff --git a/net/atm/common.c b/net/atm/common.c
index 940404a73b3d..1b9c52a02cd3 100644
--- a/net/atm/common.c
+++ b/net/atm/common.c
@@ -792,7 +792,7 @@ int vcc_getsockopt(struct socket *sock, int level, int optname,
        default:
                if (level == SOL_SOCKET)
                        return -EINVAL;
-                        break;
+                break;
        }
        if (!vcc->dev || !vcc->dev->ops->getsockopt)
                return -EINVAL;
diff --git a/net/atm/lec.c b/net/atm/lec.c
index d98bde1a0ac8..181d70c73d70 100644
--- a/net/atm/lec.c
+++ b/net/atm/lec.c
@@ -220,7 +220,6 @@ static unsigned char *get_tr_dst(unsigned char *packet, unsigned char *rdesc)
 static int lec_open(struct net_device *dev)
 {
        netif_start_queue(dev);
-        memset(&dev->stats, 0, sizeof(struct net_device_stats));
        return 0;
 }
diff --git a/net/ax25/af_ax25.c b/net/ax25/af_ax25.c
index cfdfd7e2a172..26eaebf4aaa9 100644
--- a/net/ax25/af_ax25.c
+++ b/net/ax25/af_ax25.c
@@ -1103,7 +1103,7 @@ done:
 out:
        release_sock(sk);
-        return 0;
+        return err;
 }
 /*
diff --git a/net/ax25/ax25_route.c b/net/ax25/ax25_route.c
index 7805945a5fd6..a1690845dc6e 100644
--- a/net/ax25/ax25_route.c
+++ b/net/ax25/ax25_route.c
@@ -412,7 +412,7 @@ int ax25_rt_autobind(ax25_cb *ax25, ax25_address *addr)
 {
        ax25_uid_assoc *user;
        ax25_route *ax25_rt;
-        int err;
+        int err = 0;
        if ((ax25_rt = ax25_get_route(addr, NULL)) == NULL)
                return -EHOSTUNREACH;
@@ -453,7 +453,7 @@ int ax25_rt_autobind(ax25_cb *ax25, ax25_address *addr)
 put:
        ax25_put_route(ax25_rt);
-        return 0;
+        return err;
 }
 struct sk_buff *ax25_rt_build_path(struct sk_buff *skb, ax25_address *src,
diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c
index 421c45bd1b95..ed0f22f57668 100644
--- a/net/bluetooth/af_bluetooth.c
+++ b/net/bluetooth/af_bluetooth.c
@@ -297,13 +297,12 @@ unsigned int bt_sock_poll(struct file * file, struct socket *sock, poll_table *w
                mask |= POLLERR;
        if (sk->sk_shutdown & RCV_SHUTDOWN)
-                mask |= POLLRDHUP;
+                mask |= POLLRDHUP | POLLIN | POLLRDNORM;
        if (sk->sk_shutdown == SHUTDOWN_MASK)
                mask |= POLLHUP;
-        if (!skb_queue_empty(&sk->sk_receive_queue) ||
+        if (!skb_queue_empty(&sk->sk_receive_queue))
-                        (sk->sk_shutdown & RCV_SHUTDOWN))
                mask |= POLLIN | POLLRDNORM;
        if (sk->sk_state == BT_CLOSED)
diff --git a/net/bluetooth/rfcomm/core.c b/net/bluetooth/rfcomm/core.c
index 7dca91bb8c57..15ea84ba344e 100644
--- a/net/bluetooth/rfcomm/core.c
+++ b/net/bluetooth/rfcomm/core.c
@@ -179,13 +179,13 @@ static unsigned char rfcomm_crc_table[256] = {
 /* FCS on 2 bytes */
 static inline u8 __fcs(u8 *data)
 {
-        return (0xff - __crc(data));
+        return 0xff - __crc(data);
 }
 /* FCS on 3 bytes */
 static inline u8 __fcs2(u8 *data)
 {
-        return (0xff - rfcomm_crc_table[__crc(data) ^ data[2]]);
+        return 0xff - rfcomm_crc_table[__crc(data) ^ data[2]];
 }
 /* Check FCS */
diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index c03d2c3ff03e..89ad25a76202 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -61,30 +61,27 @@ static int port_cost(struct net_device *dev)
 }
-/*
+/* Check for port carrier transistions. */
- * Check for port carrier transistions.
- * Called from work queue to allow for calling functions that
- * might sleep (such as speed check), and to debounce.
- */
 void br_port_carrier_check(struct net_bridge_port *p)
 {
        struct net_device *dev = p->dev;
        struct net_bridge *br = p->br;
-        if (netif_carrier_ok(dev))
+        if (netif_running(dev) && netif_carrier_ok(dev))
                p->path_cost = port_cost(dev);
-        if (netif_running(br->dev)) {
+        if (!netif_running(br->dev))
-                spin_lock_bh(&br->lock);
+                return;
-                if (netif_carrier_ok(dev)) {
-                        if (p->state == BR_STATE_DISABLED)
+        spin_lock_bh(&br->lock);
-                                br_stp_enable_port(p);
+        if (netif_running(dev) && netif_carrier_ok(dev)) {
-                } else {
+                if (p->state == BR_STATE_DISABLED)
-                        if (p->state != BR_STATE_DISABLED)
+                        br_stp_enable_port(p);
-                                br_stp_disable_port(p);
+        } else {
-                }
+                if (p->state != BR_STATE_DISABLED)
-                spin_unlock_bh(&br->lock);
+                        br_stp_disable_port(p);
        }
+        spin_unlock_bh(&br->lock);
 }
 static void release_nbp(struct kobject *kobj)
diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c
index 826cd5221536..6d04cfdf4541 100644
--- a/net/bridge/br_input.c
+++ b/net/bridge/br_input.c
@@ -141,7 +141,7 @@ struct sk_buff *br_handle_frame(struct sk_buff *skb)
        const unsigned char *dest = eth_hdr(skb)->h_dest;
        int (*rhook)(struct sk_buff *skb);
-        if (skb->pkt_type == PACKET_LOOPBACK)
+        if (unlikely(skb->pkt_type == PACKET_LOOPBACK))
                return skb;
        if (!is_valid_ether_addr(eth_hdr(skb)->h_source))
diff --git a/net/bridge/br_netfilter.c b/net/bridge/br_netfilter.c
index 137f23259a93..77f7b5fda45a 100644
--- a/net/bridge/br_netfilter.c
+++ b/net/bridge/br_netfilter.c
@@ -209,6 +209,72 @@ static inline void nf_bridge_update_protocol(struct sk_buff *skb)
                skb->protocol = htons(ETH_P_PPP_SES);
 }
+/* When handing a packet over to the IP layer
+ * check whether we have a skb that is in the
+ * expected format
+ */
+int br_parse_ip_options(struct sk_buff *skb)
+{
+        struct ip_options *opt;
+        struct iphdr *iph;
+        struct net_device *dev = skb->dev;
+        u32 len;
+        iph = ip_hdr(skb);
+        opt = &(IPCB(skb)->opt);
+        /* Basic sanity checks */
+        if (iph->ihl < 5 || iph->version != 4)
+                goto inhdr_error;
+        if (!pskb_may_pull(skb, iph->ihl*4))
+                goto inhdr_error;
+        iph = ip_hdr(skb);
+        if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
+                goto inhdr_error;
+        len = ntohs(iph->tot_len);
+        if (skb->len < len) {
+                IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS);
+                goto drop;
+        } else if (len < (iph->ihl*4))
+                goto inhdr_error;
+        if (pskb_trim_rcsum(skb, len)) {
+                IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
+                goto drop;
+        }
+        /* Zero out the CB buffer if no options present */
+        if (iph->ihl == 5) {
+                memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
+                return 0;
+        }
+        opt->optlen = iph->ihl*4 - sizeof(struct iphdr);
+        if (ip_options_compile(dev_net(dev), opt, skb))
+                goto inhdr_error;
+        /* Check correct handling of SRR option */
+        if (unlikely(opt->srr)) {
+                struct in_device *in_dev = __in_dev_get_rcu(dev);
+                if (in_dev && !IN_DEV_SOURCE_ROUTE(in_dev))
+                        goto drop;
+                if (ip_options_rcv_srr(skb))
+                        goto drop;
+        }
+        return 0;
+inhdr_error:
+        IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INHDRERRORS);
+drop:
+        return -1;
+}
 /* Fill in the header for fragmented IP packets handled by
 * the IPv4 connection tracking code.
 */
@@ -549,7 +615,6 @@ static unsigned int br_nf_pre_routing(unsigned int hook, struct sk_buff *skb,
 {
        struct net_bridge_port *p;
        struct net_bridge *br;
-        struct iphdr *iph;
        __u32 len = nf_bridge_encap_header_len(skb);
        if (unlikely(!pskb_may_pull(skb, len)))
@@ -578,28 +643,9 @@ static unsigned int br_nf_pre_routing(unsigned int hook, struct sk_buff *skb,
        nf_bridge_pull_encap_header_rcsum(skb);
-        if (!pskb_may_pull(skb, sizeof(struct iphdr)))
+        if (br_parse_ip_options(skb))
-                goto inhdr_error;
+                /* Drop invalid packet */
+                goto out;
-        iph = ip_hdr(skb);
-        if (iph->ihl < 5 || iph->version != 4)
-                goto inhdr_error;
-        if (!pskb_may_pull(skb, 4 * iph->ihl))
-                goto inhdr_error;
-        iph = ip_hdr(skb);
-        if (ip_fast_csum((__u8 *) iph, iph->ihl) != 0)
-                goto inhdr_error;
-        len = ntohs(iph->tot_len);
-        if (skb->len < len || len < 4 * iph->ihl)
-                goto inhdr_error;
-        pskb_trim_rcsum(skb, len);
-        /* BUG: Should really parse the IP options here. */
-        memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
        nf_bridge_put(skb->nf_bridge);
        if (!nf_bridge_alloc(skb))
@@ -614,8 +660,6 @@ static unsigned int br_nf_pre_routing(unsigned int hook, struct sk_buff *skb,
        return NF_STOLEN;
-inhdr_error:
-//      IP_INC_STATS_BH(IpInHdrErrors);
 out:
        return NF_DROP;
 }
@@ -759,14 +803,19 @@ static unsigned int br_nf_forward_arp(unsigned int hook, struct sk_buff *skb,
 #if defined(CONFIG_NF_CONNTRACK_IPV4) || defined(CONFIG_NF_CONNTRACK_IPV4_MODULE)
 static int br_nf_dev_queue_xmit(struct sk_buff *skb)
 {
+        int ret;
        if (skb->nfct != NULL && skb->protocol == htons(ETH_P_IP) &&
            skb->len + nf_bridge_mtu_reduction(skb) > skb->dev->mtu &&
            !skb_is_gso(skb)) {
-                /* BUG: Should really parse the IP options here. */
+                if (br_parse_ip_options(skb))
-                memset(IPCB(skb), 0, sizeof(struct inet_skb_parm));
+                        /* Drop invalid packet */
-                return ip_fragment(skb, br_dev_queue_push_xmit);
+                        return NF_DROP;
+                ret = ip_fragment(skb, br_dev_queue_push_xmit);
        } else
-                return br_dev_queue_push_xmit(skb);
+                ret = br_dev_queue_push_xmit(skb);
+        return ret;
 }
 #else
 static int br_nf_dev_queue_xmit(struct sk_buff *skb)
diff --git a/net/caif/caif_dev.c b/net/caif/caif_dev.c
index 0b586e9d1378..b99369a055d1 100644
--- a/net/caif/caif_dev.c
+++ b/net/caif/caif_dev.c
@@ -9,6 +9,8 @@
 *  and Sakari Ailus <sakari.ailus@nokia.com>
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/version.h>
 #include <linux/module.h>
 #include <linux/kernel.h>
@@ -171,7 +173,7 @@ static int receive(struct sk_buff *skb, struct net_device *dev,
        net = dev_net(dev);
        pkt = cfpkt_fromnative(CAIF_DIR_IN, skb);
        caifd = caif_get(dev);
-        if (!caifd || !caifd->layer.up || !caifd->layer.up->ctrlcmd)
+        if (!caifd || !caifd->layer.up || !caifd->layer.up->receive)
                return NET_RX_DROP;
        if (caifd->layer.up->receive(caifd->layer.up, pkt))
@@ -214,7 +216,7 @@ static int caif_device_notify(struct notifier_block *me, unsigned long what,
        switch (what) {
        case NETDEV_REGISTER:
-                pr_info("CAIF: %s():register %s\n", __func__, dev->name);
+                netdev_info(dev, "register\n");
                caifd = caif_device_alloc(dev);
                if (caifd == NULL)
                        break;
@@ -225,14 +227,13 @@ static int caif_device_notify(struct notifier_block *me, unsigned long what,
                break;
        case NETDEV_UP:
-                pr_info("CAIF: %s(): up %s\n", __func__, dev->name);
+                netdev_info(dev, "up\n");
                caifd = caif_get(dev);
                if (caifd == NULL)
                        break;
                caifdev = netdev_priv(dev);
                if (atomic_read(&caifd->state) == NETDEV_UP) {
-                        pr_info("CAIF: %s():%s already up\n",
+                        netdev_info(dev, "already up\n");
-                                __func__, dev->name);
                        break;
                }
                atomic_set(&caifd->state, what);
@@ -273,7 +274,7 @@ static int caif_device_notify(struct notifier_block *me, unsigned long what,
                caifd = caif_get(dev);
                if (caifd == NULL)
                        break;
-                pr_info("CAIF: %s():going down %s\n", __func__, dev->name);
+                netdev_info(dev, "going down\n");
                if (atomic_read(&caifd->state) == NETDEV_GOING_DOWN ||
                        atomic_read(&caifd->state) == NETDEV_DOWN)
@@ -295,11 +296,10 @@ static int caif_device_notify(struct notifier_block *me, unsigned long what,
                caifd = caif_get(dev);
                if (caifd == NULL)
                        break;
-                pr_info("CAIF: %s(): down %s\n", __func__, dev->name);
+                netdev_info(dev, "down\n");
                if (atomic_read(&caifd->in_use))
-                        pr_warning("CAIF: %s(): "
+                        netdev_warn(dev,
-                                   "Unregistering an active CAIF device: %s\n",
+                                    "Unregistering an active CAIF device\n");
-                                   __func__, dev->name);
                cfcnfg_del_phy_layer(get_caif_conf(), &caifd->layer);
                dev_put(dev);
                atomic_set(&caifd->state, what);
@@ -307,7 +307,7 @@ static int caif_device_notify(struct notifier_block *me, unsigned long what,
        case NETDEV_UNREGISTER:
                caifd = caif_get(dev);
-                pr_info("CAIF: %s(): unregister %s\n", __func__, dev->name);
+                netdev_info(dev, "unregister\n");
                atomic_set(&caifd->state, what);
                caif_device_destroy(dev);
                break;
@@ -391,7 +391,7 @@ static int __init caif_device_init(void)
        int result;
        cfg = cfcnfg_create();
        if (!cfg) {
-                pr_warning("CAIF: %s(): can't create cfcnfg.\n", __func__);
+                pr_warn("can't create cfcnfg\n");
                goto err_cfcnfg_create_failed;
        }
        result = register_pernet_device(&caif_net_ops);
diff --git a/net/caif/caif_socket.c b/net/caif/caif_socket.c
index 8ce904786116..4d918f8f4e67 100644
--- a/net/caif/caif_socket.c
+++ b/net/caif/caif_socket.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/fs.h>
 #include <linux/init.h>
 #include <linux/module.h>
@@ -28,9 +30,6 @@
 MODULE_LICENSE("GPL");
 MODULE_ALIAS_NETPROTO(AF_CAIF);
-#define CAIF_DEF_SNDBUF (4096*10)
-#define CAIF_DEF_RCVBUF (4096*100)
 /*
 * CAIF state is re-using the TCP socket states.
 * caif_states stored in sk_state reflect the state as reported by
@@ -157,9 +156,7 @@ static int caif_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
        if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >=
                (unsigned)sk->sk_rcvbuf && rx_flow_is_on(cf_sk)) {
-                trace_printk("CAIF: %s():"
+                pr_debug("sending flow OFF (queue len = %d %d)\n",
-                        " sending flow OFF (queue len = %d %d)\n",
-                        __func__,
                        atomic_read(&cf_sk->sk.sk_rmem_alloc),
                        sk_rcvbuf_lowwater(cf_sk));
                set_rx_flow_off(cf_sk);
@@ -172,9 +169,7 @@ static int caif_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
                return err;
        if (!sk_rmem_schedule(sk, skb->truesize) && rx_flow_is_on(cf_sk)) {
                set_rx_flow_off(cf_sk);
-                trace_printk("CAIF: %s():"
+                pr_debug("sending flow OFF due to rmem_schedule\n");
-                        " sending flow OFF due to rmem_schedule\n",
-                        __func__);
                dbfs_atomic_inc(&cnt.num_rx_flow_off);
                caif_flow_ctrl(sk, CAIF_MODEMCMD_FLOW_OFF_REQ);
        }
@@ -275,8 +270,7 @@ static void caif_ctrl_cb(struct cflayer *layr,
                break;
        default:
-                pr_debug("CAIF: %s(): Unexpected flow command %d\n",
+                pr_debug("Unexpected flow command %d\n", flow);
-                                __func__, flow);
        }
 }
@@ -536,8 +530,7 @@ static int transmit_skb(struct sk_buff *skb, struct caifsock *cf_sk,
                /* Slight paranoia, probably not needed. */
                if (unlikely(loopcnt++ > 1000)) {
-                        pr_warning("CAIF: %s(): transmit retries failed,"
+                        pr_warn("transmit retries failed, error = %d\n", ret);
-                                " error = %d\n", __func__, ret);
                        break;
                }
@@ -902,8 +895,7 @@ static int caif_connect(struct socket *sock, struct sockaddr *uaddr,
        cf_sk->maxframe = dev->mtu - (headroom + tailroom);
        dev_put(dev);
        if (cf_sk->maxframe < 1) {
-                pr_warning("CAIF: %s(): CAIF Interface MTU too small (%d)\n",
+                pr_warn("CAIF Interface MTU too small (%d)\n", dev->mtu);
-                        __func__, dev->mtu);
                err = -ENODEV;
                goto out;
        }
@@ -1123,10 +1115,6 @@ static int caif_create(struct net *net, struct socket *sock, int protocol,
        /* Store the protocol */
        sk->sk_protocol = (unsigned char) protocol;
-        /* Sendbuf dictates the amount of outbound packets not yet sent */
-        sk->sk_sndbuf = CAIF_DEF_SNDBUF;
-        sk->sk_rcvbuf = CAIF_DEF_RCVBUF;
        /*
         * Lock in order to try to stop someone from opening the socket
         * too early.
diff --git a/net/caif/cfcnfg.c b/net/caif/cfcnfg.c
index 1c29189b344d..41adafd18914 100644
--- a/net/caif/cfcnfg.c
+++ b/net/caif/cfcnfg.c
@@ -3,6 +3,9 @@
 * Author:      Sjur Brendeland/sjur.brandeland@stericsson.com
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/kernel.h>
 #include <linux/stddef.h>
 #include <linux/slab.h>
@@ -78,7 +81,7 @@ struct cfcnfg *cfcnfg_create(void)
        /* Initiate this layer */
        this = kzalloc(sizeof(struct cfcnfg), GFP_ATOMIC);
        if (!this) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        this->mux = cfmuxl_create();
@@ -106,7 +109,7 @@ struct cfcnfg *cfcnfg_create(void)
        layer_set_up(this->ctrl, this);
        return this;
 out_of_mem:
-        pr_warning("CAIF: %s(): Out of memory\n", __func__);
+        pr_warn("Out of memory\n");
        kfree(this->mux);
        kfree(this->ctrl);
        kfree(this);
@@ -194,7 +197,7 @@ int cfcnfg_disconn_adapt_layer(struct cfcnfg *cnfg, struct cflayer *adap_layer)
        caif_assert(adap_layer != NULL);
        channel_id = adap_layer->id;
        if (adap_layer->dn == NULL || channel_id == 0) {
-                pr_err("CAIF: %s():adap_layer->id is 0\n", __func__);
+                pr_err("adap_layer->dn == NULL or adap_layer->id is 0\n");
                ret = -ENOTCONN;
                goto end;
        }
@@ -204,9 +207,8 @@ int cfcnfg_disconn_adapt_layer(struct cfcnfg *cnfg, struct cflayer *adap_layer)
        layer_set_up(servl, NULL);
        ret = cfctrl_linkdown_req(cnfg->ctrl, channel_id, adap_layer);
        if (servl == NULL) {
-                pr_err("CAIF: %s(): PROTOCOL ERROR "
+                pr_err("PROTOCOL ERROR - Error removing service_layer Channel_Id(%d)",
-                       "- Error removing service_layer Channel_Id(%d)",
+                       channel_id);
-                        __func__, channel_id);
                ret = -EINVAL;
                goto end;
        }
@@ -216,18 +218,14 @@ int cfcnfg_disconn_adapt_layer(struct cfcnfg *cnfg, struct cflayer *adap_layer)
                phyinfo = cfcnfg_get_phyinfo(cnfg, phyid);
                if (phyinfo == NULL) {
-                        pr_warning("CAIF: %s(): "
+                        pr_warn("No interface to send disconnect to\n");
-                                "No interface to send disconnect to\n",
-                                __func__);
                        ret = -ENODEV;
                        goto end;
                }
                if (phyinfo->id != phyid ||
                        phyinfo->phy_layer->id != phyid ||
                        phyinfo->frm_layer->id != phyid) {
-                        pr_err("CAIF: %s(): "
+                        pr_err("Inconsistency in phy registration\n");
-                                "Inconsistency in phy registration\n",
-                                __func__);
                        ret = -EINVAL;
                        goto end;
                }
@@ -276,21 +274,20 @@ int cfcnfg_add_adaptation_layer(struct cfcnfg *cnfg,
 {
        struct cflayer *frml;
        if (adap_layer == NULL) {
-                pr_err("CAIF: %s(): adap_layer is zero", __func__);
+                pr_err("adap_layer is zero\n");
                return -EINVAL;
        }
        if (adap_layer->receive == NULL) {
-                pr_err("CAIF: %s(): adap_layer->receive is NULL", __func__);
+                pr_err("adap_layer->receive is NULL\n");
                return -EINVAL;
        }
        if (adap_layer->ctrlcmd == NULL) {
-                pr_err("CAIF: %s(): adap_layer->ctrlcmd == NULL", __func__);
+                pr_err("adap_layer->ctrlcmd == NULL\n");
                return -EINVAL;
        }
        frml = cnfg->phy_layers[param->phyid].frm_layer;
        if (frml == NULL) {
-                pr_err("CAIF: %s(): Specified PHY type does not exist!",
+                pr_err("Specified PHY type does not exist!\n");
-                        __func__);
                return -ENODEV;
        }
        caif_assert(param->phyid == cnfg->phy_layers[param->phyid].id);
@@ -330,9 +327,7 @@ cfcnfg_linkup_rsp(struct cflayer *layer, u8 channel_id, enum cfctrl_srv serv,
        struct net_device *netdev;
        if (adapt_layer == NULL) {
-                pr_debug("CAIF: %s(): link setup response "
+                pr_debug("link setup response but no client exist, send linkdown back\n");
-                                "but no client exist, send linkdown back\n",
-                                __func__);
                cfctrl_linkdown_req(cnfg->ctrl, channel_id, NULL);
                return;
        }
@@ -374,13 +369,11 @@ cfcnfg_linkup_rsp(struct cflayer *layer, u8 channel_id, enum cfctrl_srv serv,
                servicel = cfdbgl_create(channel_id, &phyinfo->dev_info);
                break;
        default:
-                pr_err("CAIF: %s(): Protocol error. "
+                pr_err("Protocol error. Link setup response - unknown channel type\n");
-                        "Link setup response - unknown channel type\n",
-                        __func__);
                return;
        }
        if (!servicel) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        layer_set_dn(servicel, cnfg->mux);
@@ -418,7 +411,7 @@ cfcnfg_add_phy_layer(struct cfcnfg *cnfg, enum cfcnfg_phy_type phy_type,
                }
        }
        if (*phyid == 0) {
-                pr_err("CAIF: %s(): No Available PHY ID\n", __func__);
+                pr_err("No Available PHY ID\n");
                return;
        }
@@ -427,7 +420,7 @@ cfcnfg_add_phy_layer(struct cfcnfg *cnfg, enum cfcnfg_phy_type phy_type,
                phy_driver =
                    cfserl_create(CFPHYTYPE_FRAG, *phyid, stx);
                if (!phy_driver) {
-                        pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                        pr_warn("Out of memory\n");
                        return;
                }
@@ -436,7 +429,7 @@ cfcnfg_add_phy_layer(struct cfcnfg *cnfg, enum cfcnfg_phy_type phy_type,
                phy_driver = NULL;
                break;
        default:
-                pr_err("CAIF: %s(): %d", __func__, phy_type);
+                pr_err("%d\n", phy_type);
                return;
                break;
        }
@@ -455,7 +448,7 @@ cfcnfg_add_phy_layer(struct cfcnfg *cnfg, enum cfcnfg_phy_type phy_type,
        phy_layer->type = phy_type;
        frml = cffrml_create(*phyid, fcs);
        if (!frml) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        cnfg->phy_layers[*phyid].frm_layer = frml;
diff --git a/net/caif/cfctrl.c b/net/caif/cfctrl.c
index 563145fdc4c3..08f267a109aa 100644
--- a/net/caif/cfctrl.c
+++ b/net/caif/cfctrl.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -36,7 +38,7 @@ struct cflayer *cfctrl_create(void)
        struct cfctrl *this =
                kmalloc(sizeof(struct cfctrl), GFP_ATOMIC);
        if (!this) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfctrl, serv.layer) == 0);
@@ -132,9 +134,7 @@ struct cfctrl_request_info *cfctrl_remove_req(struct cfctrl *ctrl,
        list_for_each_entry_safe(p, tmp, &ctrl->list, list) {
                if (cfctrl_req_eq(req, p)) {
                        if (p != first)
-                                pr_warning("CAIF: %s(): Requests are not "
+                                pr_warn("Requests are not received in order\n");
-                                        "received in order\n",
-                                        __func__);
                        atomic_set(&ctrl->rsp_seq_no,
                                         p->sequence_no);
@@ -177,7 +177,7 @@ void cfctrl_enum_req(struct cflayer *layer, u8 physlinkid)
        int ret;
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        caif_assert(offsetof(struct cfctrl, serv.layer) == 0);
@@ -189,8 +189,7 @@ void cfctrl_enum_req(struct cflayer *layer, u8 physlinkid)
        ret =
            cfctrl->serv.layer.dn->transmit(cfctrl->serv.layer.dn, pkt);
        if (ret < 0) {
-                pr_err("CAIF: %s(): Could not transmit enum message\n",
+                pr_err("Could not transmit enum message\n");
-                        __func__);
                cfpkt_destroy(pkt);
        }
 }
@@ -208,7 +207,7 @@ int cfctrl_linkup_request(struct cflayer *layer,
        char utility_name[16];
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return -ENOMEM;
        }
        cfpkt_addbdy(pkt, CFCTRL_CMD_LINK_SETUP);
@@ -253,13 +252,13 @@ int cfctrl_linkup_request(struct cflayer *layer,
                               param->u.utility.paramlen);
                break;
        default:
-                pr_warning("CAIF: %s():Request setup of bad link type = %d\n",
+                pr_warn("Request setup of bad link type = %d\n",
-                           __func__, param->linktype);
+                        param->linktype);
                return -EINVAL;
        }
        req = kzalloc(sizeof(*req), GFP_KERNEL);
        if (!req) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return -ENOMEM;
        }
        req->client_layer = user_layer;
@@ -276,8 +275,7 @@ int cfctrl_linkup_request(struct cflayer *layer,
        ret =
            cfctrl->serv.layer.dn->transmit(cfctrl->serv.layer.dn, pkt);
        if (ret < 0) {
-                pr_err("CAIF: %s(): Could not transmit linksetup request\n",
+                pr_err("Could not transmit linksetup request\n");
-                        __func__);
                cfpkt_destroy(pkt);
                return -ENODEV;
        }
@@ -291,7 +289,7 @@ int cfctrl_linkdown_req(struct cflayer *layer, u8 channelid,
        struct cfctrl *cfctrl = container_obj(layer);
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return -ENOMEM;
        }
        cfpkt_addbdy(pkt, CFCTRL_CMD_LINK_DESTROY);
@@ -300,8 +298,7 @@ int cfctrl_linkdown_req(struct cflayer *layer, u8 channelid,
        ret =
            cfctrl->serv.layer.dn->transmit(cfctrl->serv.layer.dn, pkt);
        if (ret < 0) {
-                pr_err("CAIF: %s(): Could not transmit link-down request\n",
+                pr_err("Could not transmit link-down request\n");
-                        __func__);
                cfpkt_destroy(pkt);
        }
        return ret;
@@ -313,7 +310,7 @@ void cfctrl_sleep_req(struct cflayer *layer)
        struct cfctrl *cfctrl = container_obj(layer);
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        cfpkt_addbdy(pkt, CFCTRL_CMD_SLEEP);
@@ -330,7 +327,7 @@ void cfctrl_wake_req(struct cflayer *layer)
        struct cfctrl *cfctrl = container_obj(layer);
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        cfpkt_addbdy(pkt, CFCTRL_CMD_WAKE);
@@ -347,7 +344,7 @@ void cfctrl_getstartreason_req(struct cflayer *layer)
        struct cfctrl *cfctrl = container_obj(layer);
        struct cfpkt *pkt = cfpkt_create(CFPKT_CTRL_PKT_LEN);
        if (!pkt) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return;
        }
        cfpkt_addbdy(pkt, CFCTRL_CMD_START_REASON);
@@ -364,12 +361,11 @@ void cfctrl_cancel_req(struct cflayer *layr, struct cflayer *adap_layer)
        struct cfctrl_request_info *p, *tmp;
        struct cfctrl *ctrl = container_obj(layr);
        spin_lock(&ctrl->info_list_lock);
-        pr_warning("CAIF: %s(): enter\n", __func__);
+        pr_warn("enter\n");
        list_for_each_entry_safe(p, tmp, &ctrl->list, list) {
                if (p->client_layer == adap_layer) {
-                        pr_warning("CAIF: %s(): cancel req :%d\n", __func__,
+                        pr_warn("cancel req :%d\n", p->sequence_no);
-                                        p->sequence_no);
                        list_del(&p->list);
                        kfree(p);
                }
@@ -520,9 +516,8 @@ static int cfctrl_recv(struct cflayer *layer, struct cfpkt *pkt)
                                cfpkt_extr_head(pkt, &param, len);
                                break;
                        default:
-                                pr_warning("CAIF: %s(): Request setup "
+                                pr_warn("Request setup - invalid link type (%d)\n",
-                                           "- invalid link type (%d)",
+                                        serv);
-                                           __func__, serv);
                                goto error;
                        }
@@ -532,9 +527,7 @@ static int cfctrl_recv(struct cflayer *layer, struct cfpkt *pkt)
                        if (CFCTRL_ERR_BIT == (CFCTRL_ERR_BIT & cmdrsp) ||
                                cfpkt_erroneous(pkt)) {
-                                pr_err("CAIF: %s(): Invalid O/E bit or parse "
+                                pr_err("Invalid O/E bit or parse error on CAIF control channel\n");
-                                       "error on CAIF control channel",
-                                        __func__);
                                cfctrl->res.reject_rsp(cfctrl->serv.layer.up,
                                                       0,
                                                       req ? req->client_layer
@@ -556,8 +549,7 @@ static int cfctrl_recv(struct cflayer *layer, struct cfpkt *pkt)
                cfctrl->res.linkdestroy_rsp(cfctrl->serv.layer.up, linkid);
                break;
        case CFCTRL_CMD_LINK_ERR:
-                pr_err("CAIF: %s(): Frame Error Indication received\n",
+                pr_err("Frame Error Indication received\n");
-                        __func__);
                cfctrl->res.linkerror_ind();
                break;
        case CFCTRL_CMD_ENUM:
@@ -576,7 +568,7 @@ static int cfctrl_recv(struct cflayer *layer, struct cfpkt *pkt)
                cfctrl->res.radioset_rsp();
                break;
        default:
-                pr_err("CAIF: %s(): Unrecognized Control Frame\n", __func__);
+                pr_err("Unrecognized Control Frame\n");
                goto error;
                break;
        }
@@ -595,8 +587,7 @@ static void cfctrl_ctrlcmd(struct cflayer *layr, enum caif_ctrlcmd ctrl,
        case CAIF_CTRLCMD_FLOW_OFF_IND:
                spin_lock(&this->info_list_lock);
                if (!list_empty(&this->list)) {
-                        pr_debug("CAIF: %s(): Received flow off in "
+                        pr_debug("Received flow off in control layer\n");
-                                   "control layer", __func__);
                }
                spin_unlock(&this->info_list_lock);
                break;
@@ -620,7 +611,7 @@ static int handle_loop(struct cfctrl *ctrl, int cmd, struct cfpkt *pkt)
                        if (!ctrl->loop_linkused[linkid])
                                goto found;
                spin_unlock(&ctrl->loop_linkid_lock);
-                pr_err("CAIF: %s(): Out of link-ids\n", __func__);
+                pr_err("Out of link-ids\n");
                return -EINVAL;
 found:
                if (!ctrl->loop_linkused[linkid])
diff --git a/net/caif/cfdbgl.c b/net/caif/cfdbgl.c
index 676648cac8dd..496fda9ac66f 100644
--- a/net/caif/cfdbgl.c
+++ b/net/caif/cfdbgl.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/slab.h>
 #include <net/caif/caif_layer.h>
@@ -17,7 +19,7 @@ struct cflayer *cfdbgl_create(u8 channel_id, struct dev_info *dev_info)
 {
        struct cfsrvl *dbg = kmalloc(sizeof(struct cfsrvl), GFP_ATOMIC);
        if (!dbg) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfsrvl, layer) == 0);
diff --git a/net/caif/cfdgml.c b/net/caif/cfdgml.c
index ed9d53aff280..d3ed264ad6c4 100644
--- a/net/caif/cfdgml.c
+++ b/net/caif/cfdgml.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -26,7 +28,7 @@ struct cflayer *cfdgml_create(u8 channel_id, struct dev_info *dev_info)
 {
        struct cfsrvl *dgm = kmalloc(sizeof(struct cfsrvl), GFP_ATOMIC);
        if (!dgm) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfsrvl, layer) == 0);
@@ -49,14 +51,14 @@ static int cfdgml_receive(struct cflayer *layr, struct cfpkt *pkt)
        caif_assert(layr->ctrlcmd != NULL);
        if (cfpkt_extr_head(pkt, &cmd, 1) < 0) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
        if ((cmd & DGM_CMD_BIT) == 0) {
                if (cfpkt_extr_head(pkt, &dgmhdr, 3) < 0) {
-                        pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                        pr_err("Packet is erroneous!\n");
                        cfpkt_destroy(pkt);
                        return -EPROTO;
                }
@@ -75,8 +77,7 @@ static int cfdgml_receive(struct cflayer *layr, struct cfpkt *pkt)
                return 0;
        default:
                cfpkt_destroy(pkt);
-                pr_info("CAIF: %s(): Unknown datagram control %d (0x%x)\n",
+                pr_info("Unknown datagram control %d (0x%x)\n", cmd, cmd);
-                        __func__, cmd, cmd);
                return -EPROTO;
        }
 }
diff --git a/net/caif/cffrml.c b/net/caif/cffrml.c
index e86a4ca3b217..a445043931ae 100644
--- a/net/caif/cffrml.c
+++ b/net/caif/cffrml.c
@@ -6,6 +6,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -32,7 +34,7 @@ struct cflayer *cffrml_create(u16 phyid, bool use_fcs)
 {
        struct cffrml *this = kmalloc(sizeof(struct cffrml), GFP_ATOMIC);
        if (!this) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cffrml, layer) == 0);
@@ -83,7 +85,7 @@ static int cffrml_receive(struct cflayer *layr, struct cfpkt *pkt)
        if (cfpkt_setlen(pkt, len) < 0) {
                ++cffrml_rcv_error;
-                pr_err("CAIF: %s():Framing length error (%d)\n", __func__, len);
+                pr_err("Framing length error (%d)\n", len);
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -99,14 +101,14 @@ static int cffrml_receive(struct cflayer *layr, struct cfpkt *pkt)
                        cfpkt_add_trail(pkt, &tmp, 2);
                        ++cffrml_rcv_error;
                        ++cffrml_rcv_checsum_error;
-                        pr_info("CAIF: %s(): Frame checksum error "
+                        pr_info("Frame checksum error (0x%x != 0x%x)\n",
-                                "(0x%x != 0x%x)\n", __func__, hdrchks, pktchks);
+                                hdrchks, pktchks);
                        return -EILSEQ;
                }
        }
        if (cfpkt_erroneous(pkt)) {
                ++cffrml_rcv_error;
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -132,7 +134,7 @@ static int cffrml_transmit(struct cflayer *layr, struct cfpkt *pkt)
        cfpkt_add_head(pkt, &tmp, 2);
        cfpkt_info(pkt)->hdr_len += 2;
        if (cfpkt_erroneous(pkt)) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                return -EPROTO;
        }
        ret = layr->dn->transmit(layr->dn, pkt);
diff --git a/net/caif/cfmuxl.c b/net/caif/cfmuxl.c
index 80c8d332b258..46f34b2e0478 100644
--- a/net/caif/cfmuxl.c
+++ b/net/caif/cfmuxl.c
@@ -3,6 +3,9 @@
 * Author:      Sjur Brendeland/sjur.brandeland@stericsson.com
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -190,7 +193,7 @@ static int cfmuxl_receive(struct cflayer *layr, struct cfpkt *pkt)
        u8 id;
        struct cflayer *up;
        if (cfpkt_extr_head(pkt, &id, 1) < 0) {
-                pr_err("CAIF: %s(): erroneous Caif Packet\n", __func__);
+                pr_err("erroneous Caif Packet\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -199,8 +202,8 @@ static int cfmuxl_receive(struct cflayer *layr, struct cfpkt *pkt)
        up = get_up(muxl, id);
        spin_unlock(&muxl->receive_lock);
        if (up == NULL) {
-                pr_info("CAIF: %s():Received data on unknown link ID = %d "
+                pr_info("Received data on unknown link ID = %d (0x%x) up == NULL",
-                        "(0x%x)  up == NULL", __func__, id, id);
+                        id, id);
                cfpkt_destroy(pkt);
                /*
                 * Don't return ERROR, since modem misbehaves and sends out
@@ -223,9 +226,8 @@ static int cfmuxl_transmit(struct cflayer *layr, struct cfpkt *pkt)
        struct caif_payload_info *info = cfpkt_info(pkt);
        dn = get_dn(muxl, cfpkt_info(pkt)->dev_info);
        if (dn == NULL) {
-                pr_warning("CAIF: %s(): Send data on unknown phy "
+                pr_warn("Send data on unknown phy ID = %d (0x%x)\n",
-                           "ID = %d (0x%x)\n",
+                        info->dev_info->id, info->dev_info->id);
-                           __func__, info->dev_info->id, info->dev_info->id);
                return -ENOTCONN;
        }
        info->hdr_len += 1;
diff --git a/net/caif/cfpkt_skbuff.c b/net/caif/cfpkt_skbuff.c
index c49a6695793a..d7e865e2ff65 100644
--- a/net/caif/cfpkt_skbuff.c
+++ b/net/caif/cfpkt_skbuff.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/string.h>
 #include <linux/skbuff.h>
 #include <linux/hardirq.h>
@@ -12,11 +14,12 @@
 #define PKT_PREFIX  48
 #define PKT_POSTFIX 2
 #define PKT_LEN_WHEN_EXTENDING 128
-#define PKT_ERROR(pkt, errmsg) do {        \
+#define PKT_ERROR(pkt, errmsg)             \
-    cfpkt_priv(pkt)->erronous = true;      \
+do {                                       \
-    skb_reset_tail_pointer(&pkt->skb);     \
+        cfpkt_priv(pkt)->erronous = true;  \
-    pr_warning("CAIF: " errmsg);\
+        skb_reset_tail_pointer(&pkt->skb); \
-  } while (0)
+        pr_warn(errmsg);                   \
+} while (0)
 struct cfpktq {
        struct sk_buff_head head;
@@ -130,13 +133,13 @@ int cfpkt_extr_head(struct cfpkt *pkt, void *data, u16 len)
                return -EPROTO;
        if (unlikely(len > skb->len)) {
-                PKT_ERROR(pkt, "cfpkt_extr_head read beyond end of packet\n");
+                PKT_ERROR(pkt, "read beyond end of packet\n");
                return -EPROTO;
        }
        if (unlikely(len > skb_headlen(skb))) {
                if (unlikely(skb_linearize(skb) != 0)) {
-                        PKT_ERROR(pkt, "cfpkt_extr_head linearize failed\n");
+                        PKT_ERROR(pkt, "linearize failed\n");
                        return -EPROTO;
                }
        }
@@ -156,11 +159,11 @@ int cfpkt_extr_trail(struct cfpkt *pkt, void *dta, u16 len)
                return -EPROTO;
        if (unlikely(skb_linearize(skb) != 0)) {
-                PKT_ERROR(pkt, "cfpkt_extr_trail linearize failed\n");
+                PKT_ERROR(pkt, "linearize failed\n");
                return -EPROTO;
        }
        if (unlikely(skb->data + len > skb_tail_pointer(skb))) {
-                PKT_ERROR(pkt, "cfpkt_extr_trail read beyond end of packet\n");
+                PKT_ERROR(pkt, "read beyond end of packet\n");
                return -EPROTO;
        }
        from = skb_tail_pointer(skb) - len;
@@ -202,7 +205,7 @@ int cfpkt_add_body(struct cfpkt *pkt, const void *data, u16 len)
                /* Make sure data is writable */
                if (unlikely(skb_cow_data(skb, addlen, &lastskb) < 0)) {
-                        PKT_ERROR(pkt, "cfpkt_add_body: cow failed\n");
+                        PKT_ERROR(pkt, "cow failed\n");
                        return -EPROTO;
                }
                /*
@@ -211,8 +214,7 @@ int cfpkt_add_body(struct cfpkt *pkt, const void *data, u16 len)
                 * lengths of the top SKB.
                 */
                if (lastskb != skb) {
-                        pr_warning("CAIF: %s(): Packet is non-linear\n",
+                        pr_warn("Packet is non-linear\n");
-                                   __func__);
                        skb->len += len;
                        skb->data_len += len;
                }
@@ -242,14 +244,14 @@ int cfpkt_add_head(struct cfpkt *pkt, const void *data2, u16 len)
        if (unlikely(is_erronous(pkt)))
                return -EPROTO;
        if (unlikely(skb_headroom(skb) < len)) {
-                PKT_ERROR(pkt, "cfpkt_add_head: no headroom\n");
+                PKT_ERROR(pkt, "no headroom\n");
                return -EPROTO;
        }
        /* Make sure data is writable */
        ret = skb_cow_data(skb, 0, &lastskb);
        if (unlikely(ret < 0)) {
-                PKT_ERROR(pkt, "cfpkt_add_head: cow failed\n");
+                PKT_ERROR(pkt, "cow failed\n");
                return ret;
        }
@@ -283,7 +285,7 @@ inline u16 cfpkt_iterate(struct cfpkt *pkt,
        if (unlikely(is_erronous(pkt)))
                return -EPROTO;
        if (unlikely(skb_linearize(&pkt->skb) != 0)) {
-                PKT_ERROR(pkt, "cfpkt_iterate: linearize failed\n");
+                PKT_ERROR(pkt, "linearize failed\n");
                return -EPROTO;
        }
        return iter_func(data, pkt->skb.data, cfpkt_getlen(pkt));
@@ -309,7 +311,7 @@ int cfpkt_setlen(struct cfpkt *pkt, u16 len)
        /* Need to expand SKB */
        if (unlikely(!cfpkt_pad_trail(pkt, len - skb->len)))
-                PKT_ERROR(pkt, "cfpkt_setlen: skb_pad_trail failed\n");
+                PKT_ERROR(pkt, "skb_pad_trail failed\n");
        return cfpkt_getlen(pkt);
 }
@@ -380,8 +382,7 @@ struct cfpkt *cfpkt_split(struct cfpkt *pkt, u16 pos)
                return NULL;
        if (skb->data + pos > skb_tail_pointer(skb)) {
-                PKT_ERROR(pkt,
+                PKT_ERROR(pkt, "trying to split beyond end of packet\n");
-                          "cfpkt_split: trying to split beyond end of packet");
                return NULL;
        }
@@ -455,17 +456,17 @@ int cfpkt_raw_append(struct cfpkt *pkt, void **buf, unsigned int buflen)
                return -EPROTO;
        /* Make sure SKB is writable */
        if (unlikely(skb_cow_data(skb, 0, &lastskb) < 0)) {
-                PKT_ERROR(pkt, "cfpkt_raw_append: skb_cow_data failed\n");
+                PKT_ERROR(pkt, "skb_cow_data failed\n");
                return -EPROTO;
        }
        if (unlikely(skb_linearize(skb) != 0)) {
-                PKT_ERROR(pkt, "cfpkt_raw_append: linearize failed\n");
+                PKT_ERROR(pkt, "linearize failed\n");
                return -EPROTO;
        }
        if (unlikely(skb_tailroom(skb) < buflen)) {
-                PKT_ERROR(pkt, "cfpkt_raw_append: buffer too short - failed\n");
+                PKT_ERROR(pkt, "buffer too short - failed\n");
                return -EPROTO;
        }
@@ -483,14 +484,13 @@ int cfpkt_raw_extract(struct cfpkt *pkt, void **buf, unsigned int buflen)
                return -EPROTO;
        if (unlikely(buflen > skb->len)) {
-                PKT_ERROR(pkt, "cfpkt_raw_extract: buflen too large "
+                PKT_ERROR(pkt, "buflen too large - failed\n");
-                                "- failed\n");
                return -EPROTO;
        }
        if (unlikely(buflen > skb_headlen(skb))) {
                if (unlikely(skb_linearize(skb) != 0)) {
-                        PKT_ERROR(pkt, "cfpkt_raw_extract: linearize failed\n");
+                        PKT_ERROR(pkt, "linearize failed\n");
                        return -EPROTO;
                }
        }
diff --git a/net/caif/cfrfml.c b/net/caif/cfrfml.c
index 9a699242d104..bde8481e8d25 100644
--- a/net/caif/cfrfml.c
+++ b/net/caif/cfrfml.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -48,7 +50,7 @@ struct cflayer *cfrfml_create(u8 channel_id, struct dev_info *dev_info,
                kzalloc(sizeof(struct cfrfml), GFP_ATOMIC);
        if (!this) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
@@ -178,9 +180,7 @@ out:
                        cfpkt_destroy(rfml->incomplete_frm);
                rfml->incomplete_frm = NULL;
-                pr_info("CAIF: %s(): "
+                pr_info("Connection error %d triggered on RFM link\n", err);
-                                "Connection error %d triggered on RFM link\n",
-                                __func__, err);
                /* Trigger connection error upon failure.*/
                layr->up->ctrlcmd(layr->up, CAIF_CTRLCMD_REMOTE_SHUTDOWN_IND,
@@ -280,9 +280,7 @@ static int cfrfml_transmit(struct cflayer *layr, struct cfpkt *pkt)
 out:
        if (err != 0) {
-                pr_info("CAIF: %s(): "
+                pr_info("Connection error %d triggered on RFM link\n", err);
-                                "Connection error %d triggered on RFM link\n",
-                                __func__, err);
                /* Trigger connection error upon failure.*/
                layr->up->ctrlcmd(layr->up, CAIF_CTRLCMD_REMOTE_SHUTDOWN_IND,
diff --git a/net/caif/cfserl.c b/net/caif/cfserl.c
index a11fbd68a13d..9297f7dea9d8 100644
--- a/net/caif/cfserl.c
+++ b/net/caif/cfserl.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
@@ -34,7 +36,7 @@ struct cflayer *cfserl_create(int type, int instance, bool use_stx)
 {
        struct cfserl *this = kmalloc(sizeof(struct cfserl), GFP_ATOMIC);
        if (!this) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfserl, layer) == 0);
diff --git a/net/caif/cfsrvl.c b/net/caif/cfsrvl.c
index f40939a91211..ab5e542526bf 100644
--- a/net/caif/cfsrvl.c
+++ b/net/caif/cfsrvl.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/errno.h>
@@ -79,8 +81,7 @@ static void cfservl_ctrlcmd(struct cflayer *layr, enum caif_ctrlcmd ctrl,
                layr->up->ctrlcmd(layr->up, ctrl, phyid);
                break;
        default:
-                pr_warning("CAIF: %s(): "
+                pr_warn("Unexpected ctrl in cfsrvl (%d)\n", ctrl);
-                           "Unexpected ctrl in cfsrvl (%d)\n", __func__, ctrl);
                /* We have both modem and phy flow on, send flow on */
                layr->up->ctrlcmd(layr->up, ctrl, phyid);
                service->phy_flow_on = true;
@@ -107,14 +108,12 @@ static int cfservl_modemcmd(struct cflayer *layr, enum caif_modemcmd ctrl)
                        u8 flow_on = SRVL_FLOW_ON;
                        pkt = cfpkt_create(SRVL_CTRL_PKT_SIZE);
                        if (!pkt) {
-                                pr_warning("CAIF: %s(): Out of memory\n",
+                                pr_warn("Out of memory\n");
-                                        __func__);
                                return -ENOMEM;
                        }
                        if (cfpkt_add_head(pkt, &flow_on, 1) < 0) {
-                                pr_err("CAIF: %s(): Packet is erroneous!\n",
+                                pr_err("Packet is erroneous!\n");
-                                        __func__);
                                cfpkt_destroy(pkt);
                                return -EPROTO;
                        }
@@ -131,14 +130,12 @@ static int cfservl_modemcmd(struct cflayer *layr, enum caif_modemcmd ctrl)
                        u8 flow_off = SRVL_FLOW_OFF;
                        pkt = cfpkt_create(SRVL_CTRL_PKT_SIZE);
                        if (!pkt) {
-                                pr_warning("CAIF: %s(): Out of memory\n",
+                                pr_warn("Out of memory\n");
-                                        __func__);
                                return -ENOMEM;
                        }
                        if (cfpkt_add_head(pkt, &flow_off, 1) < 0) {
-                                pr_err("CAIF: %s(): Packet is erroneous!\n",
+                                pr_err("Packet is erroneous!\n");
-                                        __func__);
                                cfpkt_destroy(pkt);
                                return -EPROTO;
                        }
diff --git a/net/caif/cfutill.c b/net/caif/cfutill.c
index 02795aff57a4..efad410e4c82 100644
--- a/net/caif/cfutill.c
+++ b/net/caif/cfutill.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/slab.h>
@@ -26,7 +28,7 @@ struct cflayer *cfutill_create(u8 channel_id, struct dev_info *dev_info)
 {
        struct cfsrvl *util = kmalloc(sizeof(struct cfsrvl), GFP_ATOMIC);
        if (!util) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfsrvl, layer) == 0);
@@ -47,7 +49,7 @@ static int cfutill_receive(struct cflayer *layr, struct cfpkt *pkt)
        caif_assert(layr->up->receive != NULL);
        caif_assert(layr->up->ctrlcmd != NULL);
        if (cfpkt_extr_head(pkt, &cmd, 1) < 0) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -64,16 +66,14 @@ static int cfutill_receive(struct cflayer *layr, struct cfpkt *pkt)
                cfpkt_destroy(pkt);
                return 0;
        case UTIL_REMOTE_SHUTDOWN:      /* Remote Shutdown Request */
-                pr_err("CAIF: %s(): REMOTE SHUTDOWN REQUEST RECEIVED\n",
+                pr_err("REMOTE SHUTDOWN REQUEST RECEIVED\n");
-                        __func__);
                layr->ctrlcmd(layr, CAIF_CTRLCMD_REMOTE_SHUTDOWN_IND, 0);
                service->open = false;
                cfpkt_destroy(pkt);
                return 0;
        default:
                cfpkt_destroy(pkt);
-                pr_warning("CAIF: %s(): Unknown service control %d (0x%x)\n",
+                pr_warn("Unknown service control %d (0x%x)\n", cmd, cmd);
-                           __func__, cmd, cmd);
                return -EPROTO;
        }
 }
diff --git a/net/caif/cfveil.c b/net/caif/cfveil.c
index 77cc09faac9a..3b425b189a99 100644
--- a/net/caif/cfveil.c
+++ b/net/caif/cfveil.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/stddef.h>
 #include <linux/slab.h>
 #include <net/caif/caif_layer.h>
@@ -25,7 +27,7 @@ struct cflayer *cfvei_create(u8 channel_id, struct dev_info *dev_info)
 {
        struct cfsrvl *vei = kmalloc(sizeof(struct cfsrvl), GFP_ATOMIC);
        if (!vei) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfsrvl, layer) == 0);
@@ -47,7 +49,7 @@ static int cfvei_receive(struct cflayer *layr, struct cfpkt *pkt)
        if (cfpkt_extr_head(pkt, &cmd, 1) < 0) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -67,8 +69,7 @@ static int cfvei_receive(struct cflayer *layr, struct cfpkt *pkt)
                cfpkt_destroy(pkt);
                return 0;
        default:                /* SET RS232 PIN */
-                pr_warning("CAIF: %s():Unknown VEI control packet %d (0x%x)!\n",
+                pr_warn("Unknown VEI control packet %d (0x%x)!\n", cmd, cmd);
-                           __func__, cmd, cmd);
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
@@ -86,7 +87,7 @@ static int cfvei_transmit(struct cflayer *layr, struct cfpkt *pkt)
        caif_assert(layr->dn->transmit != NULL);
        if (cfpkt_add_head(pkt, &tmp, 1) < 0) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                return -EPROTO;
        }
diff --git a/net/caif/cfvidl.c b/net/caif/cfvidl.c
index ada6ee2d48f5..bf6fef2a0eff 100644
--- a/net/caif/cfvidl.c
+++ b/net/caif/cfvidl.c
@@ -4,6 +4,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/slab.h>
@@ -21,7 +23,7 @@ struct cflayer *cfvidl_create(u8 channel_id, struct dev_info *dev_info)
 {
        struct cfsrvl *vid = kmalloc(sizeof(struct cfsrvl), GFP_ATOMIC);
        if (!vid) {
-                pr_warning("CAIF: %s(): Out of memory\n", __func__);
+                pr_warn("Out of memory\n");
                return NULL;
        }
        caif_assert(offsetof(struct cfsrvl, layer) == 0);
@@ -38,7 +40,7 @@ static int cfvidl_receive(struct cflayer *layr, struct cfpkt *pkt)
 {
        u32 videoheader;
        if (cfpkt_extr_head(pkt, &videoheader, 4) < 0) {
-                pr_err("CAIF: %s(): Packet is erroneous!\n", __func__);
+                pr_err("Packet is erroneous!\n");
                cfpkt_destroy(pkt);
                return -EPROTO;
        }
diff --git a/net/caif/chnl_net.c b/net/caif/chnl_net.c
index 4293e190ec53..84a422c98941 100644
--- a/net/caif/chnl_net.c
+++ b/net/caif/chnl_net.c
@@ -5,6 +5,8 @@
 * License terms: GNU General Public License (GPL) version 2
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ":%s(): " fmt, __func__
 #include <linux/version.h>
 #include <linux/fs.h>
 #include <linux/init.h>
@@ -28,9 +30,6 @@
 #define CONNECT_TIMEOUT (5 * HZ)
 #define CAIF_NET_DEFAULT_QUEUE_LEN 500
-#undef pr_debug
-#define pr_debug pr_warning
 /*This list is protected by the rtnl lock. */
 static LIST_HEAD(chnl_net_list);
@@ -142,8 +141,7 @@ static void chnl_flowctrl_cb(struct cflayer *layr, enum caif_ctrlcmd flow,
                                int phyid)
 {
        struct chnl_net *priv = container_of(layr, struct chnl_net, chnl);
-        pr_debug("CAIF: %s(): NET flowctrl func called flow: %s\n",
+        pr_debug("NET flowctrl func called flow: %s\n",
-                __func__,
                flow == CAIF_CTRLCMD_FLOW_ON_IND ? "ON" :
                flow == CAIF_CTRLCMD_INIT_RSP ? "INIT" :
                flow == CAIF_CTRLCMD_FLOW_OFF_IND ? "OFF" :
@@ -196,12 +194,12 @@ static int chnl_net_start_xmit(struct sk_buff *skb, struct net_device *dev)
        priv = netdev_priv(dev);
        if (skb->len > priv->netdev->mtu) {
-                pr_warning("CAIF: %s(): Size of skb exceeded MTU\n", __func__);
+                pr_warn("Size of skb exceeded MTU\n");
                return -ENOSPC;
        }
        if (!priv->flowenabled) {
-                pr_debug("CAIF: %s(): dropping packets flow off\n", __func__);
+                pr_debug("dropping packets flow off\n");
                return NETDEV_TX_BUSY;
        }
@@ -237,7 +235,7 @@ static int chnl_net_open(struct net_device *dev)
        ASSERT_RTNL();
        priv = netdev_priv(dev);
        if (!priv) {
-                pr_debug("CAIF: %s(): chnl_net_open: no priv\n", __func__);
+                pr_debug("chnl_net_open: no priv\n");
                return -ENODEV;
        }
@@ -246,18 +244,17 @@ static int chnl_net_open(struct net_device *dev)
                result = caif_connect_client(&priv->conn_req, &priv->chnl,
                                        &llifindex, &headroom, &tailroom);
                if (result != 0) {
-                                pr_debug("CAIF: %s(): err: "
+                                pr_debug("err: "
-                                        "Unable to register and open device,"
+                                         "Unable to register and open device,"
-                                        " Err:%d\n",
+                                         " Err:%d\n",
-                                        __func__,
+                                         result);
-                                        result);
                                goto error;
                }
                lldev = dev_get_by_index(dev_net(dev), llifindex);
                if (lldev == NULL) {
-                        pr_debug("CAIF: %s(): no interface?\n", __func__);
+                        pr_debug("no interface?\n");
                        result = -ENODEV;
                        goto error;
                }
@@ -279,9 +276,7 @@ static int chnl_net_open(struct net_device *dev)
                dev_put(lldev);
                if (mtu < 100) {
-                        pr_warning("CAIF: %s(): "
+                        pr_warn("CAIF Interface MTU too small (%d)\n", mtu);
-                                "CAIF Interface MTU too small (%d)\n",
-                                __func__, mtu);
                        result = -ENODEV;
                        goto error;
                }
@@ -296,33 +291,32 @@ static int chnl_net_open(struct net_device *dev)
        rtnl_lock();
        if (result == -ERESTARTSYS) {
-                pr_debug("CAIF: %s(): wait_event_interruptible"
+                pr_debug("wait_event_interruptible woken by a signal\n");
-                         " woken by a signal\n", __func__);
                result = -ERESTARTSYS;
                goto error;
        }
        if (result == 0) {
-                pr_debug("CAIF: %s(): connect timeout\n", __func__);
+                pr_debug("connect timeout\n");
                caif_disconnect_client(&priv->chnl);
                priv->state = CAIF_DISCONNECTED;
-                pr_debug("CAIF: %s(): state disconnected\n", __func__);
+                pr_debug("state disconnected\n");
                result = -ETIMEDOUT;
                goto error;
        }
        if (priv->state != CAIF_CONNECTED) {
-                pr_debug("CAIF: %s(): connect failed\n", __func__);
+                pr_debug("connect failed\n");
                result = -ECONNREFUSED;
                goto error;
        }
-        pr_debug("CAIF: %s(): CAIF Netdevice connected\n", __func__);
+        pr_debug("CAIF Netdevice connected\n");
        return 0;
 error:
        caif_disconnect_client(&priv->chnl);
        priv->state = CAIF_DISCONNECTED;
-        pr_debug("CAIF: %s(): state disconnected\n", __func__);
+        pr_debug("state disconnected\n");
        return result;
 }
@@ -413,7 +407,7 @@ static void caif_netlink_parms(struct nlattr *data[],
                                struct caif_connect_request *conn_req)
 {
        if (!data) {
-                pr_warning("CAIF: %s: no params data found\n", __func__);
+                pr_warn("no params data found\n");
                return;
        }
        if (data[IFLA_CAIF_IPV4_CONNID])
@@ -442,8 +436,7 @@ static int ipcaif_newlink(struct net *src_net, struct net_device *dev,
        ret = register_netdevice(dev);
        if (ret)
-                pr_warning("CAIF: %s(): device rtml registration failed\n",
+                pr_warn("device rtml registration failed\n");
-                           __func__);
        return ret;
 }
diff --git a/net/can/raw.c b/net/can/raw.c
index a10e3338f084..7d77e67e57af 100644
--- a/net/can/raw.c
+++ b/net/can/raw.c
@@ -647,12 +647,12 @@ static int raw_sendmsg(struct kiocb *iocb, struct socket *sock,
        err = memcpy_fromiovec(skb_put(skb, size), msg->msg_iov, size);
        if (err < 0)
                goto free_skb;
-        err = sock_tx_timestamp(msg, sk, skb_tx(skb));
+        err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
        if (err < 0)
                goto free_skb;
        /* to be able to check the received tx sock reference in raw_rcv() */
-        skb_tx(skb)->prevent_sk_orphan = 1;
+        skb_shinfo(skb)->tx_flags |= SKBTX_DRV_NEEDS_SK_REF;
        skb->dev = dev;
        skb->sk  = sk;
diff --git a/net/core/datagram.c b/net/core/datagram.c
index 251997a95483..4df1b7a6c1bf 100644
--- a/net/core/datagram.c
+++ b/net/core/datagram.c
@@ -746,13 +746,12 @@ unsigned int datagram_poll(struct file *file, struct socket *sock,
        if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
                mask |= POLLERR;
        if (sk->sk_shutdown & RCV_SHUTDOWN)
-                mask |= POLLRDHUP;
+                mask |= POLLRDHUP | POLLIN | POLLRDNORM;
        if (sk->sk_shutdown == SHUTDOWN_MASK)
                mask |= POLLHUP;
        /* readable? */
-        if (!skb_queue_empty(&sk->sk_receive_queue) ||
+        if (!skb_queue_empty(&sk->sk_receive_queue))
-            (sk->sk_shutdown & RCV_SHUTDOWN))
                mask |= POLLIN | POLLRDNORM;
        /* Connection-based need to check for termination and startup */
diff --git a/net/core/dev.c b/net/core/dev.c
index 660dd41aaaa6..a313bab1b754 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -129,6 +129,7 @@
 #include <linux/random.h>
 #include <trace/events/napi.h>
 #include <linux/pci.h>
+#include <linux/inetdevice.h>
 #include "net-sysfs.h"
@@ -371,6 +372,14 @@ static inline void netdev_set_addr_lockdep_class(struct net_device *dev)
 *                                                      --ANK (980803)
 */
+static inline struct list_head *ptype_head(const struct packet_type *pt)
+{
+        if (pt->type == htons(ETH_P_ALL))
+                return &ptype_all;
+        else
+                return &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK];
+}
 /**
 *      dev_add_pack - add packet handler
 *      @pt: packet type declaration
@@ -386,16 +395,11 @@ static inline void netdev_set_addr_lockdep_class(struct net_device *dev)
 void dev_add_pack(struct packet_type *pt)
 {
-        int hash;
+        struct list_head *head = ptype_head(pt);
-        spin_lock_bh(&ptype_lock);
+        spin_lock(&ptype_lock);
-        if (pt->type == htons(ETH_P_ALL))
+        list_add_rcu(&pt->list, head);
-                list_add_rcu(&pt->list, &ptype_all);
+        spin_unlock(&ptype_lock);
-        else {
-                hash = ntohs(pt->type) & PTYPE_HASH_MASK;
-                list_add_rcu(&pt->list, &ptype_base[hash]);
-        }
-        spin_unlock_bh(&ptype_lock);
 }
 EXPORT_SYMBOL(dev_add_pack);
@@ -414,15 +418,10 @@ EXPORT_SYMBOL(dev_add_pack);
 */
 void __dev_remove_pack(struct packet_type *pt)
 {
-        struct list_head *head;
+        struct list_head *head = ptype_head(pt);
        struct packet_type *pt1;
-        spin_lock_bh(&ptype_lock);
+        spin_lock(&ptype_lock);
-        if (pt->type == htons(ETH_P_ALL))
-                head = &ptype_all;
-        else
-                head = &ptype_base[ntohs(pt->type) & PTYPE_HASH_MASK];
        list_for_each_entry(pt1, head, list) {
                if (pt == pt1) {
@@ -433,7 +432,7 @@ void __dev_remove_pack(struct packet_type *pt)
        printk(KERN_WARNING "dev_remove_pack: %p not found.\n", pt);
 out:
-        spin_unlock_bh(&ptype_lock);
+        spin_unlock(&ptype_lock);
 }
 EXPORT_SYMBOL(__dev_remove_pack);
@@ -1568,6 +1567,41 @@ void netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq)
 }
 EXPORT_SYMBOL(netif_set_real_num_tx_queues);
+#ifdef CONFIG_RPS
+/**
+ *      netif_set_real_num_rx_queues - set actual number of RX queues used
+ *      @dev: Network device
+ *      @rxq: Actual number of RX queues
+ *
+ *      This must be called either with the rtnl_lock held or before
+ *      registration of the net device.  Returns 0 on success, or a
+ *      negative error code.  If called before registration, it also
+ *      sets the maximum number of queues, and always succeeds.
+ */
+int netif_set_real_num_rx_queues(struct net_device *dev, unsigned int rxq)
+{
+        int rc;
+        if (dev->reg_state == NETREG_REGISTERED) {
+                ASSERT_RTNL();
+                if (rxq > dev->num_rx_queues)
+                        return -EINVAL;
+                rc = net_rx_queue_update_kobjects(dev, dev->real_num_rx_queues,
+                                                  rxq);
+                if (rc)
+                        return rc;
+        } else {
+                dev->num_rx_queues = rxq;
+        }
+        dev->real_num_rx_queues = rxq;
+        return 0;
+}
+EXPORT_SYMBOL(netif_set_real_num_rx_queues);
+#endif
 static inline void __netif_reschedule(struct Qdisc *q)
 {
        struct softnet_data *sd;
@@ -1902,14 +1936,14 @@ static int dev_gso_segment(struct sk_buff *skb)
 /*
 * Try to orphan skb early, right before transmission by the device.
- * We cannot orphan skb if tx timestamp is requested, since
+ * We cannot orphan skb if tx timestamp is requested or the sk-reference
- * drivers need to call skb_tstamp_tx() to send the timestamp.
+ * is needed on driver level for other reasons, e.g. see net/can/raw.c
 */
 static inline void skb_orphan_try(struct sk_buff *skb)
 {
        struct sock *sk = skb->sk;
-        if (sk && !skb_tx(skb)->flags) {
+        if (sk && !skb_shinfo(skb)->tx_flags) {
                /* skb_tx_hash() wont be able to get sk.
                 * We copy sk_hash into skb->rxhash
                 */
@@ -1930,7 +1964,7 @@ static inline int skb_needs_linearize(struct sk_buff *skb,
                                      struct net_device *dev)
 {
        return skb_is_nonlinear(skb) &&
-               ((skb_has_frags(skb) && !(dev->features & NETIF_F_FRAGLIST)) ||
+               ((skb_has_frag_list(skb) && !(dev->features & NETIF_F_FRAGLIST)) ||
                (skb_shinfo(skb)->nr_frags && (!(dev->features & NETIF_F_SG) ||
                                              illegal_highdma(dev, skb))));
 }
@@ -2143,6 +2177,9 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q,
        return rc;
 }
+static DEFINE_PER_CPU(int, xmit_recursion);
+#define RECURSION_LIMIT 3
 /**
 *      dev_queue_xmit - transmit a buffer
 *      @skb: buffer to transmit
@@ -2208,10 +2245,15 @@ int dev_queue_xmit(struct sk_buff *skb)
                if (txq->xmit_lock_owner != cpu) {
+                        if (__this_cpu_read(xmit_recursion) > RECURSION_LIMIT)
+                                goto recursion_alert;
                        HARD_TX_LOCK(dev, txq, cpu);
                        if (!netif_tx_queue_stopped(txq)) {
+                                __this_cpu_inc(xmit_recursion);
                                rc = dev_hard_start_xmit(skb, dev, txq);
+                                __this_cpu_dec(xmit_recursion);
                                if (dev_xmit_complete(rc)) {
                                        HARD_TX_UNLOCK(dev, txq);
                                        goto out;
@@ -2223,7 +2265,9 @@ int dev_queue_xmit(struct sk_buff *skb)
                                       "queue packet!\n", dev->name);
                } else {
                        /* Recursion is detected! It is possible,
-                         * unfortunately */
+                         * unfortunately
+                         */
+recursion_alert:
                        if (net_ratelimit())
                                printk(KERN_CRIT "Dead loop on virtual device "
                                       "%s, fix it urgently!\n", dev->name);
@@ -2259,69 +2303,44 @@ static inline void ____napi_schedule(struct softnet_data *sd,
        __raise_softirq_irqoff(NET_RX_SOFTIRQ);
 }
-#ifdef CONFIG_RPS
-/* One global table that all flow-based protocols share. */
-struct rps_sock_flow_table *rps_sock_flow_table __read_mostly;
-EXPORT_SYMBOL(rps_sock_flow_table);
 /*
- * get_rps_cpu is called from netif_receive_skb and returns the target
+ * __skb_get_rxhash: calculate a flow hash based on src/dst addresses
- * CPU from the RPS map of the receiving queue for a given skb.
+ * and src/dst port numbers. Returns a non-zero hash number on success
- * rcu_read_lock must be held on entry.
+ * and 0 on failure.
 */
-static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
+__u32 __skb_get_rxhash(struct sk_buff *skb)
-                       struct rps_dev_flow **rflowp)
 {
+        int nhoff, hash = 0, poff;
        struct ipv6hdr *ip6;
        struct iphdr *ip;
-        struct netdev_rx_queue *rxqueue;
-        struct rps_map *map;
-        struct rps_dev_flow_table *flow_table;
-        struct rps_sock_flow_table *sock_flow_table;
-        int cpu = -1;
        u8 ip_proto;
-        u16 tcpu;
        u32 addr1, addr2, ihl;
        union {
                u32 v32;
                u16 v16[2];
        } ports;
-        if (skb_rx_queue_recorded(skb)) {
+        nhoff = skb_network_offset(skb);
-                u16 index = skb_get_rx_queue(skb);
-                if (unlikely(index >= dev->num_rx_queues)) {
-                        WARN_ONCE(dev->num_rx_queues > 1, "%s received packet "
-                                "on queue %u, but number of RX queues is %u\n",
-                                dev->name, index, dev->num_rx_queues);
-                        goto done;
-                }
-                rxqueue = dev->_rx + index;
-        } else
-                rxqueue = dev->_rx;
-        if (!rxqueue->rps_map && !rxqueue->rps_flow_table)
-                goto done;
-        if (skb->rxhash)
-                goto got_hash; /* Skip hash computation on packet header */
        switch (skb->protocol) {
        case __constant_htons(ETH_P_IP):
-                if (!pskb_may_pull(skb, sizeof(*ip)))
+                if (!pskb_may_pull(skb, sizeof(*ip) + nhoff))
                        goto done;
-                ip = (struct iphdr *) skb->data;
+                ip = (struct iphdr *) (skb->data + nhoff);
-                ip_proto = ip->protocol;
+                if (ip->frag_off & htons(IP_MF | IP_OFFSET))
+                        ip_proto = 0;
+                else
+                        ip_proto = ip->protocol;
                addr1 = (__force u32) ip->saddr;
                addr2 = (__force u32) ip->daddr;
                ihl = ip->ihl;
                break;
        case __constant_htons(ETH_P_IPV6):
-                if (!pskb_may_pull(skb, sizeof(*ip6)))
+                if (!pskb_may_pull(skb, sizeof(*ip6) + nhoff))
                        goto done;
-                ip6 = (struct ipv6hdr *) skb->data;
+                ip6 = (struct ipv6hdr *) (skb->data + nhoff);
                ip_proto = ip6->nexthdr;
                addr1 = (__force u32) ip6->saddr.s6_addr32[3];
                addr2 = (__force u32) ip6->daddr.s6_addr32[3];
@@ -2330,33 +2349,81 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
        default:
                goto done;
        }
-        switch (ip_proto) {
-        case IPPROTO_TCP:
+        ports.v32 = 0;
-        case IPPROTO_UDP:
+        poff = proto_ports_offset(ip_proto);
-        case IPPROTO_DCCP:
+        if (poff >= 0) {
-        case IPPROTO_ESP:
+                nhoff += ihl * 4 + poff;
-        case IPPROTO_AH:
+                if (pskb_may_pull(skb, nhoff + 4)) {
-        case IPPROTO_SCTP:
+                        ports.v32 = * (__force u32 *) (skb->data + nhoff);
-        case IPPROTO_UDPLITE:
-                if (pskb_may_pull(skb, (ihl * 4) + 4)) {
-                        ports.v32 = * (__force u32 *) (skb->data + (ihl * 4));
                        if (ports.v16[1] < ports.v16[0])
                                swap(ports.v16[0], ports.v16[1]);
-                        break;
                }
-        default:
-                ports.v32 = 0;
-                break;
        }
        /* get a consistent hash (same value on both flow directions) */
        if (addr2 < addr1)
                swap(addr1, addr2);
-        skb->rxhash = jhash_3words(addr1, addr2, ports.v32, hashrnd);
-        if (!skb->rxhash)
-                skb->rxhash = 1;
-got_hash:
+        hash = jhash_3words(addr1, addr2, ports.v32, hashrnd);
+        if (!hash)
+                hash = 1;
+done:
+        return hash;
+}
+EXPORT_SYMBOL(__skb_get_rxhash);
+#ifdef CONFIG_RPS
+/* One global table that all flow-based protocols share. */
+struct rps_sock_flow_table *rps_sock_flow_table __read_mostly;
+EXPORT_SYMBOL(rps_sock_flow_table);
+/*
+ * get_rps_cpu is called from netif_receive_skb and returns the target
+ * CPU from the RPS map of the receiving queue for a given skb.
+ * rcu_read_lock must be held on entry.
+ */
+static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb,
+                       struct rps_dev_flow **rflowp)
+{
+        struct netdev_rx_queue *rxqueue;
+        struct rps_map *map = NULL;
+        struct rps_dev_flow_table *flow_table;
+        struct rps_sock_flow_table *sock_flow_table;
+        int cpu = -1;
+        u16 tcpu;
+        if (skb_rx_queue_recorded(skb)) {
+                u16 index = skb_get_rx_queue(skb);
+                if (unlikely(index >= dev->real_num_rx_queues)) {
+                        WARN_ONCE(dev->real_num_rx_queues > 1,
+                                  "%s received packet on queue %u, but number "
+                                  "of RX queues is %u\n",
+                                  dev->name, index, dev->real_num_rx_queues);
+                        goto done;
+                }
+                rxqueue = dev->_rx + index;
+        } else
+                rxqueue = dev->_rx;
+        if (rxqueue->rps_map) {
+                map = rcu_dereference(rxqueue->rps_map);
+                if (map && map->len == 1) {
+                        tcpu = map->cpus[0];
+                        if (cpu_online(tcpu))
+                                cpu = tcpu;
+                        goto done;
+                }
+        } else if (!rxqueue->rps_flow_table) {
+                goto done;
+        }
+        skb_reset_network_header(skb);
+        if (!skb_get_rxhash(skb))
+                goto done;
        flow_table = rcu_dereference(rxqueue->rps_flow_table);
        sock_flow_table = rcu_dereference(rps_sock_flow_table);
        if (flow_table && sock_flow_table) {
@@ -2396,7 +2463,6 @@ got_hash:
                }
        }
-        map = rcu_dereference(rxqueue->rps_map);
        if (map) {
                tcpu = map->cpus[((u64) skb->rxhash * map->len) >> 32];
@@ -2654,7 +2720,7 @@ static int ing_filter(struct sk_buff *skb)
        skb->tc_verd = SET_TC_RTTL(skb->tc_verd, ttl);
        skb->tc_verd = SET_TC_AT(skb->tc_verd, AT_INGRESS);
-        rxq = &dev->rx_queue;
+        rxq = &dev->ingress_queue;
        q = rxq->qdisc;
        if (q != &noop_qdisc) {
@@ -2671,7 +2737,7 @@ static inline struct sk_buff *handle_ing(struct sk_buff *skb,
                                         struct packet_type **pt_prev,
                                         int *ret, struct net_device *orig_dev)
 {
-        if (skb->dev->rx_queue.qdisc == &noop_qdisc)
+        if (skb->dev->ingress_queue.qdisc == &noop_qdisc)
                goto out;
        if (*pt_prev) {
@@ -2828,8 +2894,8 @@ static int __netif_receive_skb(struct sk_buff *skb)
        if (!netdev_tstamp_prequeue)
                net_timestamp_check(skb);
-        if (vlan_tx_tag_present(skb) && vlan_hwaccel_do_receive(skb))
+        if (vlan_tx_tag_present(skb))
-                return NET_RX_SUCCESS;
+                vlan_hwaccel_do_receive(skb);
        /* if we've gotten here through NAPI, check netpoll */
        if (netpoll_receive_skb(skb))
@@ -3050,7 +3116,7 @@ out:
        return netif_receive_skb(skb);
 }
-static void napi_gro_flush(struct napi_struct *napi)
+inline void napi_gro_flush(struct napi_struct *napi)
 {
        struct sk_buff *skb, *next;
@@ -3063,6 +3129,7 @@ static void napi_gro_flush(struct napi_struct *napi)
        napi->gro_count = 0;
        napi->gro_list = NULL;
 }
+EXPORT_SYMBOL(napi_gro_flush);
 enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
 {
@@ -3077,7 +3144,7 @@ enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
        if (!(skb->dev->features & NETIF_F_GRO) || netpoll_rx_on(skb))
                goto normal;
-        if (skb_is_gso(skb) || skb_has_frags(skb))
+        if (skb_is_gso(skb) || skb_has_frag_list(skb))
                goto normal;
        rcu_read_lock();
@@ -3156,16 +3223,18 @@ normal:
 }
 EXPORT_SYMBOL(dev_gro_receive);
-static gro_result_t
+static inline gro_result_t
 __napi_gro_receive(struct napi_struct *napi, struct sk_buff *skb)
 {
        struct sk_buff *p;
        for (p = napi->gro_list; p; p = p->next) {
-                NAPI_GRO_CB(p)->same_flow =
+                unsigned long diffs;
-                        (p->dev == skb->dev) &&
-                        !compare_ether_header(skb_mac_header(p),
+                diffs = (unsigned long)p->dev ^ (unsigned long)skb->dev;
+                diffs |= compare_ether_header(skb_mac_header(p),
                                              skb_gro_mac_header(skb));
+                NAPI_GRO_CB(p)->same_flow = !diffs;
                NAPI_GRO_CB(p)->flush = 0;
        }
@@ -4871,7 +4940,7 @@ static void __netdev_init_queue_locks_one(struct net_device *dev,
 static void netdev_init_queue_locks(struct net_device *dev)
 {
        netdev_for_each_tx_queue(dev, __netdev_init_queue_locks_one, NULL);
-        __netdev_init_queue_locks_one(dev, &dev->rx_queue, NULL);
+        __netdev_init_queue_locks_one(dev, &dev->ingress_queue, NULL);
 }
 unsigned long netdev_fix_features(unsigned long features, const char *name)
@@ -4941,6 +5010,34 @@ void netif_stacked_transfer_operstate(const struct net_device *rootdev,
 }
 EXPORT_SYMBOL(netif_stacked_transfer_operstate);
+static int netif_alloc_rx_queues(struct net_device *dev)
+{
+#ifdef CONFIG_RPS
+        unsigned int i, count = dev->num_rx_queues;
+        if (count) {
+                struct netdev_rx_queue *rx;
+                rx = kcalloc(count, sizeof(struct netdev_rx_queue), GFP_KERNEL);
+                if (!rx) {
+                        pr_err("netdev: Unable to allocate %u rx queues.\n",
+                               count);
+                        return -ENOMEM;
+                }
+                dev->_rx = rx;
+                atomic_set(&rx->count, count);
+                /*
+                 * Set a pointer to first element in the array which holds the
+                 * reference count.
+                 */
+                for (i = 0; i < count; i++)
+                        rx[i].first = rx;
+        }
+#endif
+        return 0;
+}
 /**
 *      register_netdevice      - register a network device
 *      @dev: device to register
@@ -4978,24 +5075,10 @@ int register_netdevice(struct net_device *dev)
        dev->iflink = -1;
-#ifdef CONFIG_RPS
+        ret = netif_alloc_rx_queues(dev);
-        if (!dev->num_rx_queues) {
+        if (ret)
-                /*
+                goto out;
-                 * Allocate a single RX queue if driver never called
-                 * alloc_netdev_mq
-                 */
-                dev->_rx = kzalloc(sizeof(struct netdev_rx_queue), GFP_KERNEL);
-                if (!dev->_rx) {
-                        ret = -ENOMEM;
-                        goto out;
-                }
-                dev->_rx->first = dev->_rx;
-                atomic_set(&dev->_rx->count, 1);
-                dev->num_rx_queues = 1;
-        }
-#endif
        /* Init, if this function is available */
        if (dev->netdev_ops->ndo_init) {
                ret = dev->netdev_ops->ndo_init(dev);
@@ -5035,6 +5118,12 @@ int register_netdevice(struct net_device *dev)
        if (dev->features & NETIF_F_SG)
                dev->features |= NETIF_F_GSO;
+        /* Enable GRO and NETIF_F_HIGHDMA for vlans by default,
+         * vlan_dev_init() will do the dev->features check, so these features
+         * are enabled only if supported by underlying device.
+         */
+        dev->vlan_features |= (NETIF_F_GRO | NETIF_F_HIGHDMA);
        ret = call_netdevice_notifiers(NETDEV_POST_INIT, dev);
        ret = notifier_to_errno(ret);
        if (ret)
@@ -5264,7 +5353,7 @@ void netdev_run_todo(void)
                /* paranoia */
                BUG_ON(atomic_read(&dev->refcnt));
-                WARN_ON(dev->ip_ptr);
+                WARN_ON(rcu_dereference_raw(dev->ip_ptr));
                WARN_ON(dev->ip6_ptr);
                WARN_ON(dev->dn_ptr);
@@ -5363,7 +5452,7 @@ static void netdev_init_one_queue(struct net_device *dev,
 static void netdev_init_queues(struct net_device *dev)
 {
-        netdev_init_one_queue(dev, &dev->rx_queue, NULL);
+        netdev_init_one_queue(dev, &dev->ingress_queue, NULL);
        netdev_for_each_tx_queue(dev, netdev_init_one_queue, NULL);
        spin_lock_init(&dev->tx_global_lock);
 }
@@ -5386,10 +5475,6 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
        struct net_device *dev;
        size_t alloc_size;
        struct net_device *p;
-#ifdef CONFIG_RPS
-        struct netdev_rx_queue *rx;
-        int i;
-#endif
        BUG_ON(strlen(name) >= sizeof(dev->name));
@@ -5415,29 +5500,12 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
                goto free_p;
        }
-#ifdef CONFIG_RPS
-        rx = kcalloc(queue_count, sizeof(struct netdev_rx_queue), GFP_KERNEL);
-        if (!rx) {
-                printk(KERN_ERR "alloc_netdev: Unable to allocate "
-                       "rx queues.\n");
-                goto free_tx;
-        }
-        atomic_set(&rx->count, queue_count);
-        /*
-         * Set a pointer to first element in the array which holds the
-         * reference count.
-         */
-        for (i = 0; i < queue_count; i++)
-                rx[i].first = rx;
-#endif
        dev = PTR_ALIGN(p, NETDEV_ALIGN);
        dev->padded = (char *)dev - (char *)p;
        if (dev_addr_init(dev))
-                goto free_rx;
+                goto free_tx;
        dev_mc_init(dev);
        dev_uc_init(dev);
@@ -5449,8 +5517,8 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
        dev->real_num_tx_queues = queue_count;
 #ifdef CONFIG_RPS
-        dev->_rx = rx;
        dev->num_rx_queues = queue_count;
+        dev->real_num_rx_queues = queue_count;
 #endif
        dev->gso_max_size = GSO_MAX_SIZE;
@@ -5467,11 +5535,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
        strcpy(dev->name, name);
        return dev;
-free_rx:
-#ifdef CONFIG_RPS
-        kfree(rx);
 free_tx:
-#endif
        kfree(tx);
 free_p:
        kfree(p);
@@ -5658,6 +5722,10 @@ int dev_change_net_namespace(struct net_device *dev, struct net *net, const char
        /* Notify protocols, that we are about to destroy
           this device. They should clean all the things.
+           Note that dev->reg_state stays at NETREG_REGISTERED.
+           This is wanted because this way 8021q and macvlan know
+           the device is just moving and can keep their slaves up.
        */
        call_netdevice_notifiers(NETDEV_UNREGISTER, dev);
        call_netdevice_notifiers(NETDEV_UNREGISTER_BATCH, dev);
diff --git a/net/core/ethtool.c b/net/core/ethtool.c
index 7a85367b3c2f..7d7e572cedc7 100644
--- a/net/core/ethtool.c
+++ b/net/core/ethtool.c
@@ -19,6 +19,7 @@
 #include <linux/netdevice.h>
 #include <linux/bitops.h>
 #include <linux/uaccess.h>
+#include <linux/vmalloc.h>
 #include <linux/slab.h>
 /*
@@ -205,18 +206,24 @@ static noinline_for_stack int ethtool_get_drvinfo(struct net_device *dev,
        struct ethtool_drvinfo info;
        const struct ethtool_ops *ops = dev->ethtool_ops;
-        if (!ops->get_drvinfo)
-                return -EOPNOTSUPP;
        memset(&info, 0, sizeof(info));
        info.cmd = ETHTOOL_GDRVINFO;
-        ops->get_drvinfo(dev, &info);
+        if (ops && ops->get_drvinfo) {
+                ops->get_drvinfo(dev, &info);
+        } else if (dev->dev.parent && dev->dev.parent->driver) {
+                strlcpy(info.bus_info, dev_name(dev->dev.parent),
+                        sizeof(info.bus_info));
+                strlcpy(info.driver, dev->dev.parent->driver->name,
+                        sizeof(info.driver));
+        } else {
+                return -EOPNOTSUPP;
+        }
        /*
         * this method of obtaining string set info is deprecated;
         * Use ETHTOOL_GSSET_INFO instead.
         */
-        if (ops->get_sset_count) {
+        if (ops && ops->get_sset_count) {
                int rc;
                rc = ops->get_sset_count(dev, ETH_SS_TEST);
@@ -229,9 +236,9 @@ static noinline_for_stack int ethtool_get_drvinfo(struct net_device *dev,
                if (rc >= 0)
                        info.n_priv_flags = rc;
        }
-        if (ops->get_regs_len)
+        if (ops && ops->get_regs_len)
                info.regdump_len = ops->get_regs_len(dev);
-        if (ops->get_eeprom_len)
+        if (ops && ops->get_eeprom_len)
                info.eedump_len = ops->get_eeprom_len(dev);
        if (copy_to_user(useraddr, &info, sizeof(info)))
@@ -479,6 +486,38 @@ static void __rx_ntuple_filter_add(struct ethtool_rx_ntuple_list *list,
        list->count++;
 }
+/*
+ * ethtool does not (or did not) set masks for flow parameters that are
+ * not specified, so if both value and mask are 0 then this must be
+ * treated as equivalent to a mask with all bits set.  Implement that
+ * here rather than in drivers.
+ */
+static void rx_ntuple_fix_masks(struct ethtool_rx_ntuple_flow_spec *fs)
+{
+        struct ethtool_tcpip4_spec *entry = &fs->h_u.tcp_ip4_spec;
+        struct ethtool_tcpip4_spec *mask = &fs->m_u.tcp_ip4_spec;
+        if (fs->flow_type != TCP_V4_FLOW &&
+            fs->flow_type != UDP_V4_FLOW &&
+            fs->flow_type != SCTP_V4_FLOW)
+                return;
+        if (!(entry->ip4src | mask->ip4src))
+                mask->ip4src = htonl(0xffffffff);
+        if (!(entry->ip4dst | mask->ip4dst))
+                mask->ip4dst = htonl(0xffffffff);
+        if (!(entry->psrc | mask->psrc))
+                mask->psrc = htons(0xffff);
+        if (!(entry->pdst | mask->pdst))
+                mask->pdst = htons(0xffff);
+        if (!(entry->tos | mask->tos))
+                mask->tos = 0xff;
+        if (!(fs->vlan_tag | fs->vlan_tag_mask))
+                fs->vlan_tag_mask = 0xffff;
+        if (!(fs->data | fs->data_mask))
+                fs->data_mask = 0xffffffffffffffffULL;
+}
 static noinline_for_stack int ethtool_set_rx_ntuple(struct net_device *dev,
                                                    void __user *useraddr)
 {
@@ -493,6 +532,8 @@ static noinline_for_stack int ethtool_set_rx_ntuple(struct net_device *dev,
        if (copy_from_user(&cmd, useraddr, sizeof(cmd)))
                return -EFAULT;
+        rx_ntuple_fix_masks(&cmd.fs);
        /*
         * Cache filter in dev struct for GET operation only if
         * the underlying driver doesn't have its own GET operation, and
@@ -667,19 +708,19 @@ static int ethtool_get_rx_ntuple(struct net_device *dev, void __user *useraddr)
                        break;
                case IP_USER_FLOW:
                        sprintf(p, "\tSrc IP addr: 0x%x\n",
-                                fsc->fs.h_u.raw_ip4_spec.ip4src);
+                                fsc->fs.h_u.usr_ip4_spec.ip4src);
                        p += ETH_GSTRING_LEN;
                        num_strings++;
                        sprintf(p, "\tSrc IP mask: 0x%x\n",
-                                fsc->fs.m_u.raw_ip4_spec.ip4src);
+                                fsc->fs.m_u.usr_ip4_spec.ip4src);
                        p += ETH_GSTRING_LEN;
                        num_strings++;
                        sprintf(p, "\tDest IP addr: 0x%x\n",
-                                fsc->fs.h_u.raw_ip4_spec.ip4dst);
+                                fsc->fs.h_u.usr_ip4_spec.ip4dst);
                        p += ETH_GSTRING_LEN;
                        num_strings++;
                        sprintf(p, "\tDest IP mask: 0x%x\n",
-                                fsc->fs.m_u.raw_ip4_spec.ip4dst);
+                                fsc->fs.m_u.usr_ip4_spec.ip4dst);
                        p += ETH_GSTRING_LEN;
                        num_strings++;
                        break;
@@ -775,7 +816,7 @@ static int ethtool_get_regs(struct net_device *dev, char __user *useraddr)
        if (regs.len > reglen)
                regs.len = reglen;
-        regbuf = kmalloc(reglen, GFP_USER);
+        regbuf = vmalloc(reglen);
        if (!regbuf)
                return -ENOMEM;
@@ -790,7 +831,7 @@ static int ethtool_get_regs(struct net_device *dev, char __user *useraddr)
        ret = 0;
 out:
-        kfree(regbuf);
+        vfree(regbuf);
        return ret;
 }
@@ -1175,8 +1216,11 @@ static int ethtool_set_gro(struct net_device *dev, char __user *useraddr)
                return -EFAULT;
        if (edata.data) {
-                if (!dev->ethtool_ops->get_rx_csum ||
+                u32 rxcsum = dev->ethtool_ops->get_rx_csum ?
-                    !dev->ethtool_ops->get_rx_csum(dev))
+                                dev->ethtool_ops->get_rx_csum(dev) :
+                                ethtool_op_get_rx_csum(dev);
+                if (!rxcsum)
                        return -EINVAL;
                dev->features |= NETIF_F_GRO;
        } else
@@ -1402,14 +1446,22 @@ int dev_ethtool(struct net *net, struct ifreq *ifr)
        if (!dev || !netif_device_present(dev))
                return -ENODEV;
-        if (!dev->ethtool_ops)
-                return -EOPNOTSUPP;
        if (copy_from_user(&ethcmd, useraddr, sizeof(ethcmd)))
                return -EFAULT;
+        if (!dev->ethtool_ops) {
+                /* ETHTOOL_GDRVINFO does not require any driver support.
+                 * It is also unprivileged and does not change anything,
+                 * so we can take a shortcut to it. */
+                if (ethcmd == ETHTOOL_GDRVINFO)
+                        return ethtool_get_drvinfo(dev, useraddr);
+                else
+                        return -EOPNOTSUPP;
+        }
        /* Allow some commands to be done by anyone */
        switch (ethcmd) {
+        case ETHTOOL_GSET:
        case ETHTOOL_GDRVINFO:
        case ETHTOOL_GMSGLVL:
        case ETHTOOL_GCOALESCE:
diff --git a/net/core/fib_rules.c b/net/core/fib_rules.c
index 42e84e08a1be..332c2e31d048 100644
--- a/net/core/fib_rules.c
+++ b/net/core/fib_rules.c
@@ -182,7 +182,8 @@ static int fib_rule_match(struct fib_rule *rule, struct fib_rules_ops *ops,
 {
        int ret = 0;
-        if (rule->iifindex && (rule->iifindex != fl->iif))
+        if (rule->iifindex && (rule->iifindex != fl->iif) &&
+            !(fl->flags & FLOWI_FLAG_MATCH_ANY_IIF))
                goto out;
        if (rule->oifindex && (rule->oifindex != fl->oif))
@@ -225,9 +226,11 @@ jumped:
                        err = ops->action(rule, fl, flags, arg);
                if (err != -EAGAIN) {
-                        fib_rule_get(rule);
+                        if (likely(atomic_inc_not_zero(&rule->refcnt))) {
-                        arg->rule = rule;
+                                arg->rule = rule;
-                        goto out;
+                                goto out;
+                        }
+                        break;
                }
        }
diff --git a/net/core/filter.c b/net/core/filter.c
index 52b051f82a01..7adf50352918 100644
--- a/net/core/filter.c
+++ b/net/core/filter.c
@@ -638,10 +638,9 @@ int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk)
                return err;
        }
-        rcu_read_lock_bh();
+        old_fp = rcu_dereference_protected(sk->sk_filter,
-        old_fp = rcu_dereference_bh(sk->sk_filter);
+                                           sock_owned_by_user(sk));
        rcu_assign_pointer(sk->sk_filter, fp);
-        rcu_read_unlock_bh();
        if (old_fp)
                sk_filter_delayed_uncharge(sk, old_fp);
@@ -654,14 +653,13 @@ int sk_detach_filter(struct sock *sk)
        int ret = -ENOENT;
        struct sk_filter *filter;
-        rcu_read_lock_bh();
+        filter = rcu_dereference_protected(sk->sk_filter,
-        filter = rcu_dereference_bh(sk->sk_filter);
+                                           sock_owned_by_user(sk));
        if (filter) {
                rcu_assign_pointer(sk->sk_filter, NULL);
                sk_filter_delayed_uncharge(sk, filter);
                ret = 0;
        }
-        rcu_read_unlock_bh();
        return ret;
 }
 EXPORT_SYMBOL_GPL(sk_detach_filter);
diff --git a/net/core/flow.c b/net/core/flow.c
index f67dcbfe54ef..127c8a7ffd61 100644
--- a/net/core/flow.c
+++ b/net/core/flow.c
@@ -53,8 +53,7 @@ struct flow_flush_info {
 struct flow_cache {
        u32                             hash_shift;
-        unsigned long                   order;
+        struct flow_cache_percpu __percpu *percpu;
-        struct flow_cache_percpu        *percpu;
        struct notifier_block           hotcpu_notifier;
        int                             low_watermark;
        int                             high_watermark;
@@ -64,7 +63,7 @@ struct flow_cache {
 atomic_t flow_cache_genid = ATOMIC_INIT(0);
 EXPORT_SYMBOL(flow_cache_genid);
 static struct flow_cache flow_cache_global;
-static struct kmem_cache *flow_cachep;
+static struct kmem_cache *flow_cachep __read_mostly;
 static DEFINE_SPINLOCK(flow_cache_gc_lock);
 static LIST_HEAD(flow_cache_gc_list);
@@ -177,15 +176,11 @@ static u32 flow_hash_code(struct flow_cache *fc,
 {
        u32 *k = (u32 *) key;
-        return (jhash2(k, (sizeof(*key) / sizeof(u32)), fcp->hash_rnd)
+        return jhash2(k, (sizeof(*key) / sizeof(u32)), fcp->hash_rnd)
-                & (flow_cache_hash_size(fc) - 1));
+                & (flow_cache_hash_size(fc) - 1);
 }
-#if (BITS_PER_LONG == 64)
+typedef unsigned long flow_compare_t;
-typedef u64 flow_compare_t;
-#else
-typedef u32 flow_compare_t;
-#endif
 /* I hear what you're saying, use memcmp.  But memcmp cannot make
 * important assumptions that we can here, such as alignment and
@@ -357,62 +352,73 @@ void flow_cache_flush(void)
        put_online_cpus();
 }
-static void __init flow_cache_cpu_prepare(struct flow_cache *fc,
+static int __cpuinit flow_cache_cpu_prepare(struct flow_cache *fc, int cpu)
-                                          struct flow_cache_percpu *fcp)
 {
-        fcp->hash_table = (struct hlist_head *)
+        struct flow_cache_percpu *fcp = per_cpu_ptr(fc->percpu, cpu);
-                __get_free_pages(GFP_KERNEL|__GFP_ZERO, fc->order);
+        size_t sz = sizeof(struct hlist_head) * flow_cache_hash_size(fc);
-        if (!fcp->hash_table)
-                panic("NET: failed to allocate flow cache order %lu\n", fc->order);
-        fcp->hash_rnd_recalc = 1;
+        if (!fcp->hash_table) {
-        fcp->hash_count = 0;
+                fcp->hash_table = kzalloc_node(sz, GFP_KERNEL, cpu_to_node(cpu));
-        tasklet_init(&fcp->flush_tasklet, flow_cache_flush_tasklet, 0);
+                if (!fcp->hash_table) {
+                        pr_err("NET: failed to allocate flow cache sz %zu\n", sz);
+                        return -ENOMEM;
+                }
+                fcp->hash_rnd_recalc = 1;
+                fcp->hash_count = 0;
+                tasklet_init(&fcp->flush_tasklet, flow_cache_flush_tasklet, 0);
+        }
+        return 0;
 }
-static int flow_cache_cpu(struct notifier_block *nfb,
+static int __cpuinit flow_cache_cpu(struct notifier_block *nfb,
                          unsigned long action,
                          void *hcpu)
 {
        struct flow_cache *fc = container_of(nfb, struct flow_cache, hotcpu_notifier);
-        int cpu = (unsigned long) hcpu;
+        int res, cpu = (unsigned long) hcpu;
        struct flow_cache_percpu *fcp = per_cpu_ptr(fc->percpu, cpu);
-        if (action == CPU_DEAD || action == CPU_DEAD_FROZEN)
+        switch (action) {
+        case CPU_UP_PREPARE:
+        case CPU_UP_PREPARE_FROZEN:
+                res = flow_cache_cpu_prepare(fc, cpu);
+                if (res)
+                        return notifier_from_errno(res);
+                break;
+        case CPU_DEAD:
+        case CPU_DEAD_FROZEN:
                __flow_cache_shrink(fc, fcp, 0);
+                break;
+        }
        return NOTIFY_OK;
 }
-static int flow_cache_init(struct flow_cache *fc)
+static int __init flow_cache_init(struct flow_cache *fc)
 {
-        unsigned long order;
        int i;
        fc->hash_shift = 10;
        fc->low_watermark = 2 * flow_cache_hash_size(fc);
        fc->high_watermark = 4 * flow_cache_hash_size(fc);
-        for (order = 0;
-             (PAGE_SIZE << order) <
-                     (sizeof(struct hlist_head)*flow_cache_hash_size(fc));
-             order++)
-                /* NOTHING */;
-        fc->order = order;
        fc->percpu = alloc_percpu(struct flow_cache_percpu);
+        if (!fc->percpu)
+                return -ENOMEM;
-        setup_timer(&fc->rnd_timer, flow_cache_new_hashrnd,
+        for_each_online_cpu(i) {
-                    (unsigned long) fc);
+                if (flow_cache_cpu_prepare(fc, i))
-        fc->rnd_timer.expires = jiffies + FLOW_HASH_RND_PERIOD;
+                        return -ENOMEM;
-        add_timer(&fc->rnd_timer);
+        }
-        for_each_possible_cpu(i)
-                flow_cache_cpu_prepare(fc, per_cpu_ptr(fc->percpu, i));
        fc->hotcpu_notifier = (struct notifier_block){
                .notifier_call = flow_cache_cpu,
        };
        register_hotcpu_notifier(&fc->hotcpu_notifier);
+        setup_timer(&fc->rnd_timer, flow_cache_new_hashrnd,
+                    (unsigned long) fc);
+        fc->rnd_timer.expires = jiffies + FLOW_HASH_RND_PERIOD;
+        add_timer(&fc->rnd_timer);
        return 0;
 }
diff --git a/net/core/gen_estimator.c b/net/core/gen_estimator.c
index 6743146e4d6b..7c2373321b74 100644
--- a/net/core/gen_estimator.c
+++ b/net/core/gen_estimator.c
@@ -274,9 +274,9 @@ void gen_kill_estimator(struct gnet_stats_basic_packed *bstats,
        while ((e = gen_find_node(bstats, rate_est))) {
                rb_erase(&e->node, &est_root);
-                write_lock_bh(&est_lock);
+                write_lock(&est_lock);
                e->bstats = NULL;
-                write_unlock_bh(&est_lock);
+                write_unlock(&est_lock);
                list_del_rcu(&e->list);
                call_rcu(&e->e_rcu, __gen_kill_estimator);
diff --git a/net/core/iovec.c b/net/core/iovec.c
index e6b133b77ccb..72aceb1fe4fa 100644
--- a/net/core/iovec.c
+++ b/net/core/iovec.c
@@ -42,7 +42,9 @@ long verify_iovec(struct msghdr *m, struct iovec *iov, struct sockaddr *address,
        if (m->msg_namelen) {
                if (mode == VERIFY_READ) {
-                        err = move_addr_to_kernel(m->msg_name, m->msg_namelen,
+                        void __user *namep;
+                        namep = (void __user __force *) m->msg_name;
+                        err = move_addr_to_kernel(namep, m->msg_namelen,
                                                  address);
                        if (err < 0)
                                return err;
@@ -53,7 +55,7 @@ long verify_iovec(struct msghdr *m, struct iovec *iov, struct sockaddr *address,
        }
        size = m->msg_iovlen * sizeof(struct iovec);
-        if (copy_from_user(iov, m->msg_iov, size))
+        if (copy_from_user(iov, (void __user __force *) m->msg_iov, size))
                return -EFAULT;
        m->msg_iov = iov;
diff --git a/net/core/neighbour.c b/net/core/neighbour.c
index a4e0a7482c2b..b142a0d76072 100644
--- a/net/core/neighbour.c
+++ b/net/core/neighbour.c
@@ -122,7 +122,7 @@ static void neigh_cleanup_and_release(struct neighbour *neigh)
 unsigned long neigh_rand_reach_time(unsigned long base)
 {
-        return (base ? (net_random() % base) + (base >> 1) : 0);
+        return base ? (net_random() % base) + (base >> 1) : 0;
 }
 EXPORT_SYMBOL(neigh_rand_reach_time);
@@ -766,9 +766,9 @@ next_elt:
 static __inline__ int neigh_max_probes(struct neighbour *n)
 {
        struct neigh_parms *p = n->parms;
-        return (n->nud_state & NUD_PROBE ?
+        return (n->nud_state & NUD_PROBE) ?
                p->ucast_probes :
-                p->ucast_probes + p->app_probes + p->mcast_probes);
+                p->ucast_probes + p->app_probes + p->mcast_probes;
 }
 static void neigh_invalidate(struct neighbour *neigh)
@@ -1210,7 +1210,9 @@ int neigh_resolve_output(struct sk_buff *skb)
        if (!neigh_event_send(neigh, skb)) {
                int err;
                struct net_device *dev = neigh->dev;
-                if (dev->header_ops->cache && !dst->hh) {
+                if (dev->header_ops->cache &&
+                    !dst->hh &&
+                    !(dst->flags & DST_NOCACHE)) {
                        write_lock_bh(&neigh->lock);
                        if (!dst->hh)
                                neigh_hh_init(neigh, dst, dst->ops->protocol);
diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
index af4dfbadf2a0..fa81fd0a488f 100644
--- a/net/core/net-sysfs.c
+++ b/net/core/net-sysfs.c
@@ -515,7 +515,7 @@ static ssize_t rx_queue_attr_store(struct kobject *kobj, struct attribute *attr,
        return attribute->store(queue, attribute, buf, count);
 }
-static struct sysfs_ops rx_queue_sysfs_ops = {
+static const struct sysfs_ops rx_queue_sysfs_ops = {
        .show = rx_queue_attr_show,
        .store = rx_queue_attr_store,
 };
@@ -742,34 +742,38 @@ static int rx_queue_add_kobject(struct net_device *net, int index)
        return error;
 }
-static int rx_queue_register_kobjects(struct net_device *net)
+int
+net_rx_queue_update_kobjects(struct net_device *net, int old_num, int new_num)
 {
        int i;
        int error = 0;
-        net->queues_kset = kset_create_and_add("queues",
+        for (i = old_num; i < new_num; i++) {
-            NULL, &net->dev.kobj);
-        if (!net->queues_kset)
-                return -ENOMEM;
-        for (i = 0; i < net->num_rx_queues; i++) {
                error = rx_queue_add_kobject(net, i);
-                if (error)
+                if (error) {
+                        new_num = old_num;
                        break;
+                }
        }
-        if (error)
+        while (--i >= new_num)
-                while (--i >= 0)
+                kobject_put(&net->_rx[i].kobj);
-                        kobject_put(&net->_rx[i].kobj);
        return error;
 }
-static void rx_queue_remove_kobjects(struct net_device *net)
+static int rx_queue_register_kobjects(struct net_device *net)
 {
-        int i;
+        net->queues_kset = kset_create_and_add("queues",
+            NULL, &net->dev.kobj);
+        if (!net->queues_kset)
+                return -ENOMEM;
+        return net_rx_queue_update_kobjects(net, 0, net->real_num_rx_queues);
+}
-        for (i = 0; i < net->num_rx_queues; i++)
+static void rx_queue_remove_kobjects(struct net_device *net)
-                kobject_put(&net->_rx[i].kobj);
+{
+        net_rx_queue_update_kobjects(net, net->real_num_rx_queues, 0);
        kset_unregister(net->queues_kset);
 }
 #endif /* CONFIG_RPS */
@@ -789,12 +793,13 @@ static const void *net_netlink_ns(struct sock *sk)
        return sock_net(sk);
 }
-static struct kobj_ns_type_operations net_ns_type_operations = {
+struct kobj_ns_type_operations net_ns_type_operations = {
        .type = KOBJ_NS_TYPE_NET,
        .current_ns = net_current_ns,
        .netlink_ns = net_netlink_ns,
        .initial_ns = net_initial_ns,
 };
+EXPORT_SYMBOL_GPL(net_ns_type_operations);
 static void net_kobj_ns_exit(struct net *net)
 {
diff --git a/net/core/net-sysfs.h b/net/core/net-sysfs.h
index 805555e8b187..778e1571548d 100644
--- a/net/core/net-sysfs.h
+++ b/net/core/net-sysfs.h
@@ -4,4 +4,8 @@
 int netdev_kobject_init(void);
 int netdev_register_kobject(struct net_device *);
 void netdev_unregister_kobject(struct net_device *);
+#ifdef CONFIG_RPS
+int net_rx_queue_update_kobjects(struct net_device *, int old_num, int new_num);
+#endif
 #endif
diff --git a/net/core/pktgen.c b/net/core/pktgen.c
index 10a1ea72010d..2c0df0f95b3d 100644
--- a/net/core/pktgen.c
+++ b/net/core/pktgen.c
@@ -729,16 +729,14 @@ static int hex32_arg(const char __user *user_buffer, unsigned long maxlen,
        *num = 0;
        for (; i < maxlen; i++) {
+                int value;
                char c;
                *num <<= 4;
                if (get_user(c, &user_buffer[i]))
                        return -EFAULT;
-                if ((c >= '0') && (c <= '9'))
+                value = hex_to_bin(c);
-                        *num |= c - '0';
+                if (value >= 0)
-                else if ((c >= 'a') && (c <= 'f'))
+                        *num |= value;
-                        *num |= c - 'a' + 10;
-                else if ((c >= 'A') && (c <= 'F'))
-                        *num |= c - 'A' + 10;
                else
                        break;
        }
@@ -3907,8 +3905,6 @@ static void __exit pg_cleanup(void)
 {
        struct pktgen_thread *t;
        struct list_head *q, *n;
-        wait_queue_head_t queue;
-        init_waitqueue_head(&queue);
        /* Stop all interfaces & threads */
diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
index f78d821bd935..b2a718dfd720 100644
--- a/net/core/rtnetlink.c
+++ b/net/core/rtnetlink.c
@@ -612,36 +612,7 @@ static void copy_rtnl_link_stats(struct rtnl_link_stats *a,
 static void copy_rtnl_link_stats64(void *v, const struct rtnl_link_stats64 *b)
 {
-        struct rtnl_link_stats64 a;
+        memcpy(v, b, sizeof(*b));
-        a.rx_packets = b->rx_packets;
-        a.tx_packets = b->tx_packets;
-        a.rx_bytes = b->rx_bytes;
-        a.tx_bytes = b->tx_bytes;
-        a.rx_errors = b->rx_errors;
-        a.tx_errors = b->tx_errors;
-        a.rx_dropped = b->rx_dropped;
-        a.tx_dropped = b->tx_dropped;
-        a.multicast = b->multicast;
-        a.collisions = b->collisions;
-        a.rx_length_errors = b->rx_length_errors;
-        a.rx_over_errors = b->rx_over_errors;
-        a.rx_crc_errors = b->rx_crc_errors;
-        a.rx_frame_errors = b->rx_frame_errors;
-        a.rx_fifo_errors = b->rx_fifo_errors;
-        a.rx_missed_errors = b->rx_missed_errors;
-        a.tx_aborted_errors = b->tx_aborted_errors;
-        a.tx_carrier_errors = b->tx_carrier_errors;
-        a.tx_fifo_errors = b->tx_fifo_errors;
-        a.tx_heartbeat_errors = b->tx_heartbeat_errors;
-        a.tx_window_errors = b->tx_window_errors;
-        a.rx_compressed = b->rx_compressed;
-        a.tx_compressed = b->tx_compressed;
-        memcpy(v, &a, sizeof(a));
 }
 /* All VF info */
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index c83b421341c0..752c1972b3a7 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -202,8 +202,6 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
        skb->data = data;
        skb_reset_tail_pointer(skb);
        skb->end = skb->tail + size;
-        kmemcheck_annotate_bitfield(skb, flags1);
-        kmemcheck_annotate_bitfield(skb, flags2);
 #ifdef NET_SKBUFF_DATA_USES_OFFSET
        skb->mac_header = ~0U;
 #endif
@@ -340,7 +338,7 @@ static void skb_release_data(struct sk_buff *skb)
                                put_page(skb_shinfo(skb)->frags[i].page);
                }
-                if (skb_has_frags(skb))
+                if (skb_has_frag_list(skb))
                        skb_drop_fraglist(skb);
                kfree(skb->head);
@@ -685,16 +683,10 @@ static void copy_skb_header(struct sk_buff *new, const struct sk_buff *old)
 struct sk_buff *skb_copy(const struct sk_buff *skb, gfp_t gfp_mask)
 {
-        int headerlen = skb->data - skb->head;
+        int headerlen = skb_headroom(skb);
-        /*
+        unsigned int size = (skb_end_pointer(skb) - skb->head) + skb->data_len;
-         *      Allocate the copy buffer
+        struct sk_buff *n = alloc_skb(size, gfp_mask);
-         */
-        struct sk_buff *n;
-#ifdef NET_SKBUFF_DATA_USES_OFFSET
-        n = alloc_skb(skb->end + skb->data_len, gfp_mask);
-#else
-        n = alloc_skb(skb->end - skb->head + skb->data_len, gfp_mask);
-#endif
        if (!n)
                return NULL;
@@ -726,20 +718,14 @@ EXPORT_SYMBOL(skb_copy);
 struct sk_buff *pskb_copy(struct sk_buff *skb, gfp_t gfp_mask)
 {
-        /*
+        unsigned int size = skb_end_pointer(skb) - skb->head;
-         *      Allocate the copy buffer
+        struct sk_buff *n = alloc_skb(size, gfp_mask);
-         */
-        struct sk_buff *n;
-#ifdef NET_SKBUFF_DATA_USES_OFFSET
-        n = alloc_skb(skb->end, gfp_mask);
-#else
-        n = alloc_skb(skb->end - skb->head, gfp_mask);
-#endif
        if (!n)
                goto out;
        /* Set the data pointer */
-        skb_reserve(n, skb->data - skb->head);
+        skb_reserve(n, skb_headroom(skb));
        /* Set the tail pointer and length */
        skb_put(n, skb_headlen(skb));
        /* Copy the bytes */
@@ -759,7 +745,7 @@ struct sk_buff *pskb_copy(struct sk_buff *skb, gfp_t gfp_mask)
                skb_shinfo(n)->nr_frags = i;
        }
-        if (skb_has_frags(skb)) {
+        if (skb_has_frag_list(skb)) {
                skb_shinfo(n)->frag_list = skb_shinfo(skb)->frag_list;
                skb_clone_fraglist(n);
        }
@@ -791,12 +777,9 @@ int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail,
 {
        int i;
        u8 *data;
-#ifdef NET_SKBUFF_DATA_USES_OFFSET
+        int size = nhead + (skb_end_pointer(skb) - skb->head) + ntail;
-        int size = nhead + skb->end + ntail;
-#else
-        int size = nhead + (skb->end - skb->head) + ntail;
-#endif
        long off;
+        bool fastpath;
        BUG_ON(nhead < 0);
@@ -810,23 +793,36 @@ int pskb_expand_head(struct sk_buff *skb, int nhead, int ntail,
                goto nodata;
        /* Copy only real data... and, alas, header. This should be
-         * optimized for the cases when header is void. */
+         * optimized for the cases when header is void.
-#ifdef NET_SKBUFF_DATA_USES_OFFSET
+         */
-        memcpy(data + nhead, skb->head, skb->tail);
+        memcpy(data + nhead, skb->head, skb_tail_pointer(skb) - skb->head);
-#else
-        memcpy(data + nhead, skb->head, skb->tail - skb->head);
+        memcpy((struct skb_shared_info *)(data + size),
-#endif
+               skb_shinfo(skb),
-        memcpy(data + size, skb_end_pointer(skb),
               offsetof(struct skb_shared_info, frags[skb_shinfo(skb)->nr_frags]));
-        for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
+        /* Check if we can avoid taking references on fragments if we own
-                get_page(skb_shinfo(skb)->frags[i].page);
+         * the last reference on skb->head. (see skb_release_data())
+         */
+        if (!skb->cloned)
+                fastpath = true;
+        else {
+                int delta = skb->nohdr ? (1 << SKB_DATAREF_SHIFT) + 1 : 1;
-        if (skb_has_frags(skb))
+                fastpath = atomic_read(&skb_shinfo(skb)->dataref) == delta;
-                skb_clone_fraglist(skb);
+        }
-        skb_release_data(skb);
+        if (fastpath) {
+                kfree(skb->head);
+        } else {
+                for (i = 0; i < skb_shinfo(skb)->nr_frags; i++)
+                        get_page(skb_shinfo(skb)->frags[i].page);
+                if (skb_has_frag_list(skb))
+                        skb_clone_fraglist(skb);
+                skb_release_data(skb);
+        }
        off = (data + nhead) - skb->head;
        skb->head     = data;
@@ -1099,7 +1095,7 @@ drop_pages:
                for (; i < nfrags; i++)
                        put_page(skb_shinfo(skb)->frags[i].page);
-                if (skb_has_frags(skb))
+                if (skb_has_frag_list(skb))
                        skb_drop_fraglist(skb);
                goto done;
        }
@@ -1194,7 +1190,7 @@ unsigned char *__pskb_pull_tail(struct sk_buff *skb, int delta)
        /* Optimization: no fragments, no reasons to preestimate
         * size of pulled pages. Superb.
         */
-        if (!skb_has_frags(skb))
+        if (!skb_has_frag_list(skb))
                goto pull_pages;
        /* Estimate size of pulled pages. */
@@ -2323,7 +2319,7 @@ next_skb:
                st->frag_data = NULL;
        }
-        if (st->root_skb == st->cur_skb && skb_has_frags(st->root_skb)) {
+        if (st->root_skb == st->cur_skb && skb_has_frag_list(st->root_skb)) {
                st->cur_skb = skb_shinfo(st->root_skb)->frag_list;
                st->frag_idx = 0;
                goto next_skb;
@@ -2893,7 +2889,7 @@ int skb_cow_data(struct sk_buff *skb, int tailbits, struct sk_buff **trailer)
                return -ENOMEM;
        /* Easy case. Most of packets will go this way. */
-        if (!skb_has_frags(skb)) {
+        if (!skb_has_frag_list(skb)) {
                /* A little of trouble, not enough of space for trailer.
                 * This should not happen, when stack is tuned to generate
                 * good frames. OK, on miss we reallocate and reserve even more
@@ -2928,7 +2924,7 @@ int skb_cow_data(struct sk_buff *skb, int tailbits, struct sk_buff **trailer)
                if (skb1->next == NULL && tailbits) {
                        if (skb_shinfo(skb1)->nr_frags ||
-                            skb_has_frags(skb1) ||
+                            skb_has_frag_list(skb1) ||
                            skb_tailroom(skb1) < tailbits)
                                ntail = tailbits + 128;
                }
@@ -2937,7 +2933,7 @@ int skb_cow_data(struct sk_buff *skb, int tailbits, struct sk_buff **trailer)
                    skb_cloned(skb1) ||
                    ntail ||
                    skb_shinfo(skb1)->nr_frags ||
-                    skb_has_frags(skb1)) {
+                    skb_has_frag_list(skb1)) {
                        struct sk_buff *skb2;
                        /* Fuck, we are miserable poor guys... */
@@ -3020,7 +3016,7 @@ void skb_tstamp_tx(struct sk_buff *orig_skb,
        } else {
                /*
                 * no hardware time stamps available,
-                 * so keep the skb_shared_tx and only
+                 * so keep the shared tx_flags and only
                 * store software time stamp
                 */
                skb->tstamp = ktime_get_real();
diff --git a/net/core/sock.c b/net/core/sock.c
index ef30e9d286e7..42365deeba27 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1557,6 +1557,8 @@ struct sk_buff *sock_alloc_send_skb(struct sock *sk, unsigned long size,
 EXPORT_SYMBOL(sock_alloc_send_skb);
 static void __lock_sock(struct sock *sk)
+        __releases(&sk->sk_lock.slock)
+        __acquires(&sk->sk_lock.slock)
 {
        DEFINE_WAIT(wait);
@@ -1573,6 +1575,8 @@ static void __lock_sock(struct sock *sk)
 }
 static void __release_sock(struct sock *sk)
+        __releases(&sk->sk_lock.slock)
+        __acquires(&sk->sk_lock.slock)
 {
        struct sk_buff *skb = sk->sk_backlog.head;
diff --git a/net/core/utils.c b/net/core/utils.c
index f41854470539..5fea0ab21902 100644
--- a/net/core/utils.c
+++ b/net/core/utils.c
@@ -75,7 +75,7 @@ __be32 in_aton(const char *str)
                                str++;
                }
        }
-        return(htonl(l));
+        return htonl(l);
 }
 EXPORT_SYMBOL(in_aton);
@@ -92,18 +92,19 @@ EXPORT_SYMBOL(in_aton);
 static inline int xdigit2bin(char c, int delim)
 {
+        int val;
        if (c == delim || c == '\0')
                return IN6PTON_DELIM;
        if (c == ':')
                return IN6PTON_COLON_MASK;
        if (c == '.')
                return IN6PTON_DOT;
-        if (c >= '0' && c <= '9')
-                return (IN6PTON_XDIGIT | IN6PTON_DIGIT| (c - '0'));
+        val = hex_to_bin(c);
-        if (c >= 'a' && c <= 'f')
+        if (val >= 0)
-                return (IN6PTON_XDIGIT | (c - 'a' + 10));
+                return val | IN6PTON_XDIGIT | (val < 10 ? IN6PTON_DIGIT : 0);
-        if (c >= 'A' && c <= 'F')
-                return (IN6PTON_XDIGIT | (c - 'A' + 10));
        if (delim == -1)
                return IN6PTON_DELIM;
        return IN6PTON_UNKNOWN;
diff --git a/net/dccp/ccid.h b/net/dccp/ccid.h
index 6df6f8ac9636..6d16a9070ff0 100644
--- a/net/dccp/ccid.h
+++ b/net/dccp/ccid.h
@@ -62,18 +62,14 @@ struct ccid_operations {
        void            (*ccid_hc_tx_exit)(struct sock *sk);
        void            (*ccid_hc_rx_packet_recv)(struct sock *sk,
                                                  struct sk_buff *skb);
-        int             (*ccid_hc_rx_parse_options)(struct sock *sk,
+        int             (*ccid_hc_rx_parse_options)(struct sock *sk, u8 pkt,
-                                                    unsigned char option,
+                                                    u8 opt, u8 *val, u8 len);
-                                                    unsigned char len, u16 idx,
-                                                    unsigned char* value);
        int             (*ccid_hc_rx_insert_options)(struct sock *sk,
                                                     struct sk_buff *skb);
        void            (*ccid_hc_tx_packet_recv)(struct sock *sk,
                                                  struct sk_buff *skb);
-        int             (*ccid_hc_tx_parse_options)(struct sock *sk,
+        int             (*ccid_hc_tx_parse_options)(struct sock *sk, u8 pkt,
-                                                    unsigned char option,
+                                                    u8 opt, u8 *val, u8 len);
-                                                    unsigned char len, u16 idx,
-                                                    unsigned char* value);
        int             (*ccid_hc_tx_send_packet)(struct sock *sk,
                                                  struct sk_buff *skb);
        void            (*ccid_hc_tx_packet_sent)(struct sock *sk,
@@ -168,27 +164,31 @@ static inline void ccid_hc_tx_packet_recv(struct ccid *ccid, struct sock *sk,
                ccid->ccid_ops->ccid_hc_tx_packet_recv(sk, skb);
 }
+/**
+ * ccid_hc_tx_parse_options  -  Parse CCID-specific options sent by the receiver
+ * @pkt: type of packet that @opt appears on (RFC 4340, 5.1)
+ * @opt: the CCID-specific option type (RFC 4340, 5.8 and 10.3)
+ * @val: value of @opt
+ * @len: length of @val in bytes
+ */
 static inline int ccid_hc_tx_parse_options(struct ccid *ccid, struct sock *sk,
-                                           unsigned char option,
+                                           u8 pkt, u8 opt, u8 *val, u8 len)
-                                           unsigned char len, u16 idx,
-                                           unsigned char* value)
 {
-        int rc = 0;
+        if (ccid->ccid_ops->ccid_hc_tx_parse_options == NULL)
-        if (ccid->ccid_ops->ccid_hc_tx_parse_options != NULL)
+                return 0;
-                rc = ccid->ccid_ops->ccid_hc_tx_parse_options(sk, option, len, idx,
+        return ccid->ccid_ops->ccid_hc_tx_parse_options(sk, pkt, opt, val, len);
-                                                    value);
-        return rc;
 }
+/**
+ * ccid_hc_rx_parse_options  -  Parse CCID-specific options sent by the sender
+ * Arguments are analogous to ccid_hc_tx_parse_options()
+ */
 static inline int ccid_hc_rx_parse_options(struct ccid *ccid, struct sock *sk,
-                                           unsigned char option,
+                                           u8 pkt, u8 opt, u8 *val, u8 len)
-                                           unsigned char len, u16 idx,
-                                           unsigned char* value)
 {
-        int rc = 0;
+        if (ccid->ccid_ops->ccid_hc_rx_parse_options == NULL)
-        if (ccid->ccid_ops->ccid_hc_rx_parse_options != NULL)
+                return 0;
-                rc = ccid->ccid_ops->ccid_hc_rx_parse_options(sk, option, len, idx, value);
+        return ccid->ccid_ops->ccid_hc_rx_parse_options(sk, pkt, opt, val, len);
-        return rc;
 }
 static inline int ccid_hc_rx_insert_options(struct ccid *ccid, struct sock *sk,
diff --git a/net/dccp/ccids/Kconfig b/net/dccp/ccids/Kconfig
index 8408398cd44e..0581143cb800 100644
--- a/net/dccp/ccids/Kconfig
+++ b/net/dccp/ccids/Kconfig
@@ -47,37 +47,6 @@ config IP_DCCP_CCID3_DEBUG
          If in doubt, say N.
-config IP_DCCP_CCID3_RTO
-          int "Use higher bound for nofeedback timer"
-          default 100
-          depends on IP_DCCP_CCID3 && EXPERIMENTAL
-          ---help---
-            Use higher lower bound for nofeedback timer expiration.
-            The TFRC nofeedback timer normally expires after the maximum of 4
-            RTTs and twice the current send interval (RFC 3448, 4.3). On LANs
-            with a small RTT this can mean a high processing load and reduced
-            performance, since then the nofeedback timer is triggered very
-            frequently.
-            This option enables to set a higher lower bound for the nofeedback
-            value. Values in units of milliseconds can be set here.
-            A value of 0 disables this feature by enforcing the value specified
-            in RFC 3448. The following values have been suggested as bounds for
-            experimental use:
-                * 16-20ms to match the typical multimedia inter-frame interval
-                * 100ms as a reasonable compromise [default]
-                * 1000ms corresponds to the lower TCP RTO bound (RFC 2988, 2.4)
-            The default of 100ms is a compromise between a large value for
-            efficient DCCP implementations, and a small value to avoid disrupting
-            the network in times of congestion.
-            The purpose of the nofeedback timer is to slow DCCP down when there
-            is serious network congestion: experimenting with larger values should
-            therefore not be performed on WANs.
 config IP_DCCP_TFRC_LIB
        def_bool y if IP_DCCP_CCID3
diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c
index 9b3ae9922be1..dc18172b1e59 100644
--- a/net/dccp/ccids/ccid2.c
+++ b/net/dccp/ccids/ccid2.c
@@ -25,59 +25,14 @@
 */
 #include <linux/slab.h>
 #include "../feat.h"
-#include "../ccid.h"
-#include "../dccp.h"
 #include "ccid2.h"
 #ifdef CONFIG_IP_DCCP_CCID2_DEBUG
 static int ccid2_debug;
 #define ccid2_pr_debug(format, a...)    DCCP_PR_DEBUG(ccid2_debug, format, ##a)
-static void ccid2_hc_tx_check_sanity(const struct ccid2_hc_tx_sock *hc)
-{
-        int len = 0;
-        int pipe = 0;
-        struct ccid2_seq *seqp = hc->tx_seqh;
-        /* there is data in the chain */
-        if (seqp != hc->tx_seqt) {
-                seqp = seqp->ccid2s_prev;
-                len++;
-                if (!seqp->ccid2s_acked)
-                        pipe++;
-                while (seqp != hc->tx_seqt) {
-                        struct ccid2_seq *prev = seqp->ccid2s_prev;
-                        len++;
-                        if (!prev->ccid2s_acked)
-                                pipe++;
-                        /* packets are sent sequentially */
-                        BUG_ON(dccp_delta_seqno(seqp->ccid2s_seq,
-                                                prev->ccid2s_seq ) >= 0);
-                        BUG_ON(time_before(seqp->ccid2s_sent,
-                                           prev->ccid2s_sent));
-                        seqp = prev;
-                }
-        }
-        BUG_ON(pipe != hc->tx_pipe);
-        ccid2_pr_debug("len of chain=%d\n", len);
-        do {
-                seqp = seqp->ccid2s_prev;
-                len++;
-        } while (seqp != hc->tx_seqh);
-        ccid2_pr_debug("total len=%d\n", len);
-        BUG_ON(len != hc->tx_seqbufc * CCID2_SEQBUF_LEN);
-}
 #else
 #define ccid2_pr_debug(format, a...)
-#define ccid2_hc_tx_check_sanity(hc)
 #endif
 static int ccid2_hc_tx_alloc_seq(struct ccid2_hc_tx_sock *hc)
@@ -156,19 +111,10 @@ static void ccid2_change_l_ack_ratio(struct sock *sk, u32 val)
        dp->dccps_l_ack_ratio = val;
 }
-static void ccid2_change_srtt(struct ccid2_hc_tx_sock *hc, long val)
-{
-        ccid2_pr_debug("change SRTT to %ld\n", val);
-        hc->tx_srtt = val;
-}
-static void ccid2_start_rto_timer(struct sock *sk);
 static void ccid2_hc_tx_rto_expire(unsigned long data)
 {
        struct sock *sk = (struct sock *)data;
        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
-        long s;
        bh_lock_sock(sk);
        if (sock_owned_by_user(sk)) {
@@ -178,23 +124,19 @@ static void ccid2_hc_tx_rto_expire(unsigned long data)
        ccid2_pr_debug("RTO_EXPIRE\n");
-        ccid2_hc_tx_check_sanity(hc);
        /* back-off timer */
        hc->tx_rto <<= 1;
+        if (hc->tx_rto > DCCP_RTO_MAX)
+                hc->tx_rto = DCCP_RTO_MAX;
-        s = hc->tx_rto / HZ;
+        sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto);
-        if (s > 60)
-                hc->tx_rto = 60 * HZ;
-        ccid2_start_rto_timer(sk);
        /* adjust pipe, cwnd etc */
        hc->tx_ssthresh = hc->tx_cwnd / 2;
        if (hc->tx_ssthresh < 2)
                hc->tx_ssthresh = 2;
-        hc->tx_cwnd      = 1;
+        hc->tx_cwnd     = 1;
-        hc->tx_pipe      = 0;
+        hc->tx_pipe     = 0;
        /* clear state about stuff we sent */
        hc->tx_seqt = hc->tx_seqh;
@@ -204,22 +146,11 @@ static void ccid2_hc_tx_rto_expire(unsigned long data)
        hc->tx_rpseq    = 0;
        hc->tx_rpdupack = -1;
        ccid2_change_l_ack_ratio(sk, 1);
-        ccid2_hc_tx_check_sanity(hc);
 out:
        bh_unlock_sock(sk);
        sock_put(sk);
 }
-static void ccid2_start_rto_timer(struct sock *sk)
-{
-        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
-        ccid2_pr_debug("setting RTO timeout=%ld\n", hc->tx_rto);
-        BUG_ON(timer_pending(&hc->tx_rtotimer));
-        sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto);
-}
 static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len)
 {
        struct dccp_sock *dp = dccp_sk(sk);
@@ -230,7 +161,7 @@ static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len)
        hc->tx_seqh->ccid2s_seq   = dp->dccps_gss;
        hc->tx_seqh->ccid2s_acked = 0;
-        hc->tx_seqh->ccid2s_sent  = jiffies;
+        hc->tx_seqh->ccid2s_sent  = ccid2_time_stamp;
        next = hc->tx_seqh->ccid2s_next;
        /* check if we need to alloc more space */
@@ -296,23 +227,20 @@ static void ccid2_hc_tx_packet_sent(struct sock *sk, int more, unsigned int len)
        }
 #endif
-        /* setup RTO timer */
+        sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto);
-        if (!timer_pending(&hc->tx_rtotimer))
-                ccid2_start_rto_timer(sk);
 #ifdef CONFIG_IP_DCCP_CCID2_DEBUG
        do {
                struct ccid2_seq *seqp = hc->tx_seqt;
                while (seqp != hc->tx_seqh) {
-                        ccid2_pr_debug("out seq=%llu acked=%d time=%lu\n",
+                        ccid2_pr_debug("out seq=%llu acked=%d time=%u\n",
                                       (unsigned long long)seqp->ccid2s_seq,
                                       seqp->ccid2s_acked, seqp->ccid2s_sent);
                        seqp = seqp->ccid2s_next;
                }
        } while (0);
        ccid2_pr_debug("=========\n");
-        ccid2_hc_tx_check_sanity(hc);
 #endif
 }
@@ -378,17 +306,87 @@ out_invalid_option:
        return -1;
 }
-static void ccid2_hc_tx_kill_rto_timer(struct sock *sk)
+/**
+ * ccid2_rtt_estimator - Sample RTT and compute RTO using RFC2988 algorithm
+ * This code is almost identical with TCP's tcp_rtt_estimator(), since
+ * - it has a higher sampling frequency (recommended by RFC 1323),
+ * - the RTO does not collapse into RTT due to RTTVAR going towards zero,
+ * - it is simple (cf. more complex proposals such as Eifel timer or research
+ *   which suggests that the gain should be set according to window size),
+ * - in tests it was found to work well with CCID2 [gerrit].
+ */
+static void ccid2_rtt_estimator(struct sock *sk, const long mrtt)
 {
        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
+        long m = mrtt ? : 1;
-        sk_stop_timer(sk, &hc->tx_rtotimer);
+        if (hc->tx_srtt == 0) {
-        ccid2_pr_debug("deleted RTO timer\n");
+                /* First measurement m */
+                hc->tx_srtt = m << 3;
+                hc->tx_mdev = m << 1;
+                hc->tx_mdev_max = max(hc->tx_mdev, tcp_rto_min(sk));
+                hc->tx_rttvar   = hc->tx_mdev_max;
+                hc->tx_rtt_seq  = dccp_sk(sk)->dccps_gss;
+        } else {
+                /* Update scaled SRTT as SRTT += 1/8 * (m - SRTT) */
+                m -= (hc->tx_srtt >> 3);
+                hc->tx_srtt += m;
+                /* Similarly, update scaled mdev with regard to |m| */
+                if (m < 0) {
+                        m = -m;
+                        m -= (hc->tx_mdev >> 2);
+                        /*
+                         * This neutralises RTO increase when RTT < SRTT - mdev
+                         * (see P. Sarolahti, A. Kuznetsov,"Congestion Control
+                         * in Linux TCP", USENIX 2002, pp. 49-62).
+                         */
+                        if (m > 0)
+                                m >>= 3;
+                } else {
+                        m -= (hc->tx_mdev >> 2);
+                }
+                hc->tx_mdev += m;
+                if (hc->tx_mdev > hc->tx_mdev_max) {
+                        hc->tx_mdev_max = hc->tx_mdev;
+                        if (hc->tx_mdev_max > hc->tx_rttvar)
+                                hc->tx_rttvar = hc->tx_mdev_max;
+                }
+                /*
+                 * Decay RTTVAR at most once per flight, exploiting that
+                 *  1) pipe <= cwnd <= Sequence_Window = W  (RFC 4340, 7.5.2)
+                 *  2) AWL = GSS-W+1 <= GAR <= GSS          (RFC 4340, 7.5.1)
+                 * GAR is a useful bound for FlightSize = pipe.
+                 * AWL is probably too low here, as it over-estimates pipe.
+                 */
+                if (after48(dccp_sk(sk)->dccps_gar, hc->tx_rtt_seq)) {
+                        if (hc->tx_mdev_max < hc->tx_rttvar)
+                                hc->tx_rttvar -= (hc->tx_rttvar -
+                                                  hc->tx_mdev_max) >> 2;
+                        hc->tx_rtt_seq  = dccp_sk(sk)->dccps_gss;
+                        hc->tx_mdev_max = tcp_rto_min(sk);
+                }
+        }
+        /*
+         * Set RTO from SRTT and RTTVAR
+         * As in TCP, 4 * RTTVAR >= TCP_RTO_MIN, giving a minimum RTO of 200 ms.
+         * This agrees with RFC 4341, 5:
+         *      "Because DCCP does not retransmit data, DCCP does not require
+         *       TCP's recommended minimum timeout of one second".
+         */
+        hc->tx_rto = (hc->tx_srtt >> 3) + hc->tx_rttvar;
+        if (hc->tx_rto > DCCP_RTO_MAX)
+                hc->tx_rto = DCCP_RTO_MAX;
 }
-static inline void ccid2_new_ack(struct sock *sk,
+static void ccid2_new_ack(struct sock *sk, struct ccid2_seq *seqp,
-                                 struct ccid2_seq *seqp,
+                          unsigned int *maxincr)
-                                 unsigned int *maxincr)
 {
        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
@@ -402,93 +400,27 @@ static inline void ccid2_new_ack(struct sock *sk,
                        hc->tx_cwnd += 1;
                        hc->tx_packets_acked = 0;
        }
+        /*
-        /* update RTO */
+         * FIXME: RTT is sampled several times per acknowledgment (for each
-        if (hc->tx_srtt == -1 ||
+         * entry in the Ack Vector), instead of once per Ack (as in TCP SACK).
-            time_after(jiffies, hc->tx_lastrtt + hc->tx_srtt)) {
+         * This causes the RTT to be over-estimated, since the older entries
-                unsigned long r = (long)jiffies - (long)seqp->ccid2s_sent;
+         * in the Ack Vector have earlier sending times.
-                int s;
+         * The cleanest solution is to not use the ccid2s_sent field at all
+         * and instead use DCCP timestamps: requires changes in other places.
-                /* first measurement */
+         */
-                if (hc->tx_srtt == -1) {
+        ccid2_rtt_estimator(sk, ccid2_time_stamp - seqp->ccid2s_sent);
-                        ccid2_pr_debug("R: %lu Time=%lu seq=%llu\n",
-                                       r, jiffies,
-                                       (unsigned long long)seqp->ccid2s_seq);
-                        ccid2_change_srtt(hc, r);
-                        hc->tx_rttvar = r >> 1;
-                } else {
-                        /* RTTVAR */
-                        long tmp = hc->tx_srtt - r;
-                        long srtt;
-                        if (tmp < 0)
-                                tmp *= -1;
-                        tmp >>= 2;
-                        hc->tx_rttvar *= 3;
-                        hc->tx_rttvar >>= 2;
-                        hc->tx_rttvar += tmp;
-                        /* SRTT */
-                        srtt = hc->tx_srtt;
-                        srtt *= 7;
-                        srtt >>= 3;
-                        tmp = r >> 3;
-                        srtt += tmp;
-                        ccid2_change_srtt(hc, srtt);
-                }
-                s = hc->tx_rttvar << 2;
-                /* clock granularity is 1 when based on jiffies */
-                if (!s)
-                        s = 1;
-                hc->tx_rto = hc->tx_srtt + s;
-                /* must be at least a second */
-                s = hc->tx_rto / HZ;
-                /* DCCP doesn't require this [but I like it cuz my code sux] */
-#if 1
-                if (s < 1)
-                        hc->tx_rto = HZ;
-#endif
-                /* max 60 seconds */
-                if (s > 60)
-                        hc->tx_rto = HZ * 60;
-                hc->tx_lastrtt = jiffies;
-                ccid2_pr_debug("srtt: %ld rttvar: %ld rto: %ld (HZ=%d) R=%lu\n",
-                               hc->tx_srtt, hc->tx_rttvar,
-                               hc->tx_rto, HZ, r);
-        }
-        /* we got a new ack, so re-start RTO timer */
-        ccid2_hc_tx_kill_rto_timer(sk);
-        ccid2_start_rto_timer(sk);
-}
-static void ccid2_hc_tx_dec_pipe(struct sock *sk)
-{
-        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
-        if (hc->tx_pipe == 0)
-                DCCP_BUG("pipe == 0");
-        else
-                hc->tx_pipe--;
-        if (hc->tx_pipe == 0)
-                ccid2_hc_tx_kill_rto_timer(sk);
 }
 static void ccid2_congestion_event(struct sock *sk, struct ccid2_seq *seqp)
 {
        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
-        if (time_before(seqp->ccid2s_sent, hc->tx_last_cong)) {
+        if ((s32)(seqp->ccid2s_sent - hc->tx_last_cong) < 0) {
                ccid2_pr_debug("Multiple losses in an RTT---treating as one\n");
                return;
        }
-        hc->tx_last_cong = jiffies;
+        hc->tx_last_cong = ccid2_time_stamp;
        hc->tx_cwnd      = hc->tx_cwnd / 2 ? : 1U;
        hc->tx_ssthresh  = max(hc->tx_cwnd, 2U);
@@ -510,7 +442,6 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
        int done = 0;
        unsigned int maxincr = 0;
-        ccid2_hc_tx_check_sanity(hc);
        /* check reverse path congestion */
        seqno = DCCP_SKB_CB(skb)->dccpd_seq;
@@ -620,7 +551,7 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
                                        seqp->ccid2s_acked = 1;
                                        ccid2_pr_debug("Got ack for %llu\n",
                                                       (unsigned long long)seqp->ccid2s_seq);
-                                        ccid2_hc_tx_dec_pipe(sk);
+                                        hc->tx_pipe--;
                                }
                                if (seqp == hc->tx_seqt) {
                                        done = 1;
@@ -677,7 +608,7 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
                                 * one ack vector.
                                 */
                                ccid2_congestion_event(sk, seqp);
-                                ccid2_hc_tx_dec_pipe(sk);
+                                hc->tx_pipe--;
                        }
                        if (seqp == hc->tx_seqt)
                                break;
@@ -695,7 +626,11 @@ static void ccid2_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
                hc->tx_seqt = hc->tx_seqt->ccid2s_next;
        }
-        ccid2_hc_tx_check_sanity(hc);
+        /* restart RTO timer if not all outstanding data has been acked */
+        if (hc->tx_pipe == 0)
+                sk_stop_timer(sk, &hc->tx_rtotimer);
+        else
+                sk_reset_timer(sk, &hc->tx_rtotimer, jiffies + hc->tx_rto);
 }
 static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk)
@@ -707,12 +642,8 @@ static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk)
        /* RFC 4341, 5: initialise ssthresh to arbitrarily high (max) value */
        hc->tx_ssthresh = ~0U;
-        /*
+        /* Use larger initial windows (RFC 4341, section 5). */
-         * RFC 4341, 5: "The cwnd parameter is initialized to at most four
+        hc->tx_cwnd = rfc3390_bytes_to_packets(dp->dccps_mss_cache);
-         * packets for new connections, following the rules from [RFC3390]".
-         * We need to convert the bytes of RFC3390 into the packets of RFC 4341.
-         */
-        hc->tx_cwnd = clamp(4380U / dp->dccps_mss_cache, 2U, 4U);
        /* Make sure that Ack Ratio is enabled and within bounds. */
        max_ratio = DIV_ROUND_UP(hc->tx_cwnd, 2);
@@ -723,15 +654,11 @@ static int ccid2_hc_tx_init(struct ccid *ccid, struct sock *sk)
        if (ccid2_hc_tx_alloc_seq(hc))
                return -ENOMEM;
-        hc->tx_rto       = 3 * HZ;
+        hc->tx_rto       = DCCP_TIMEOUT_INIT;
-        ccid2_change_srtt(hc, -1);
-        hc->tx_rttvar    = -1;
        hc->tx_rpdupack  = -1;
-        hc->tx_last_cong = jiffies;
+        hc->tx_last_cong = ccid2_time_stamp;
        setup_timer(&hc->tx_rtotimer, ccid2_hc_tx_rto_expire,
                        (unsigned long)sk);
-        ccid2_hc_tx_check_sanity(hc);
        return 0;
 }
@@ -740,7 +667,7 @@ static void ccid2_hc_tx_exit(struct sock *sk)
        struct ccid2_hc_tx_sock *hc = ccid2_hc_tx_sk(sk);
        int i;
-        ccid2_hc_tx_kill_rto_timer(sk);
+        sk_stop_timer(sk, &hc->tx_rtotimer);
        for (i = 0; i < hc->tx_seqbufc; i++)
                kfree(hc->tx_seqbuf[i]);
diff --git a/net/dccp/ccids/ccid2.h b/net/dccp/ccids/ccid2.h
index 1ec6a30103bb..9731c2dc1487 100644
--- a/net/dccp/ccids/ccid2.h
+++ b/net/dccp/ccids/ccid2.h
@@ -18,18 +18,23 @@
 #ifndef _DCCP_CCID2_H_
 #define _DCCP_CCID2_H_
-#include <linux/dccp.h>
 #include <linux/timer.h>
 #include <linux/types.h>
 #include "../ccid.h"
+#include "../dccp.h"
+/*
+ * CCID-2 timestamping faces the same issues as TCP timestamping.
+ * Hence we reuse/share as much of the code as possible.
+ */
+#define ccid2_time_stamp        tcp_time_stamp
 /* NUMDUPACK parameter from RFC 4341, p. 6 */
 #define NUMDUPACK       3
-struct sock;
 struct ccid2_seq {
        u64                     ccid2s_seq;
-        unsigned long           ccid2s_sent;
+        u32                     ccid2s_sent;
        int                     ccid2s_acked;
        struct ccid2_seq        *ccid2s_prev;
        struct ccid2_seq        *ccid2s_next;
@@ -42,7 +47,12 @@ struct ccid2_seq {
 * struct ccid2_hc_tx_sock - CCID2 TX half connection
 * @tx_{cwnd,ssthresh,pipe}: as per RFC 4341, section 5
 * @tx_packets_acked:        Ack counter for deriving cwnd growth (RFC 3465)
- * @tx_lastrtt:              time RTT was last measured
+ * @tx_srtt:                 smoothed RTT estimate, scaled by 2^3
+ * @tx_mdev:                 smoothed RTT variation, scaled by 2^2
+ * @tx_mdev_max:             maximum of @mdev during one flight
+ * @tx_rttvar:               moving average/maximum of @mdev_max
+ * @tx_rto:                  RTO value deriving from SRTT and RTTVAR (RFC 2988)
+ * @tx_rtt_seq:              to decay RTTVAR at most once per flight
 * @tx_rpseq:                last consecutive seqno
 * @tx_rpdupack:             dupacks since rpseq
 */
@@ -55,14 +65,19 @@ struct ccid2_hc_tx_sock {
        int                     tx_seqbufc;
        struct ccid2_seq        *tx_seqh;
        struct ccid2_seq        *tx_seqt;
-        long                    tx_rto;
-        long                    tx_srtt;
+        /* RTT measurement: variables/principles are the same as in TCP */
-        long                    tx_rttvar;
+        u32                     tx_srtt,
-        unsigned long           tx_lastrtt;
+                                tx_mdev,
+                                tx_mdev_max,
+                                tx_rttvar,
+                                tx_rto;
+        u64                     tx_rtt_seq:48;
        struct timer_list       tx_rtotimer;
        u64                     tx_rpseq;
        int                     tx_rpdupack;
-        unsigned long           tx_last_cong;
+        u32                     tx_last_cong;
        u64                     tx_high_ack;
 };
diff --git a/net/dccp/ccids/ccid3.c b/net/dccp/ccids/ccid3.c
index 95f752986497..c3f3a25bbd7a 100644
--- a/net/dccp/ccids/ccid3.c
+++ b/net/dccp/ccids/ccid3.c
@@ -54,7 +54,6 @@ static const char *ccid3_tx_state_name(enum ccid3_hc_tx_states state)
        [TFRC_SSTATE_NO_SENT]  = "NO_SENT",
        [TFRC_SSTATE_NO_FBACK] = "NO_FBACK",
        [TFRC_SSTATE_FBACK]    = "FBACK",
-        [TFRC_SSTATE_TERM]     = "TERM",
        };
        return ccid3_state_names[state];
@@ -91,19 +90,16 @@ static inline u64 rfc3390_initial_rate(struct sock *sk)
        return scaled_div(w_init << 6, hc->tx_rtt);
 }
-/*
+/**
- * Recalculate t_ipi and delta (should be called whenever X changes)
+ * ccid3_update_send_interval  -  Calculate new t_ipi = s / X_inst
+ * This respects the granularity of X_inst (64 * bytes/second).
 */
 static void ccid3_update_send_interval(struct ccid3_hc_tx_sock *hc)
 {
-        /* Calculate new t_ipi = s / X_inst (X_inst is in 64 * bytes/second) */
        hc->tx_t_ipi = scaled_div32(((u64)hc->tx_s) << 6, hc->tx_x);
-        /* Calculate new delta by delta = min(t_ipi / 2, t_gran / 2) */
+        ccid3_pr_debug("t_ipi=%u, s=%u, X=%u\n", hc->tx_t_ipi,
-        hc->tx_delta = min_t(u32, hc->tx_t_ipi / 2, TFRC_OPSYS_HALF_TIME_GRAN);
+                       hc->tx_s, (unsigned)(hc->tx_x >> 6));
-        ccid3_pr_debug("t_ipi=%u, delta=%u, s=%u, X=%u\n", hc->tx_t_ipi,
-                       hc->tx_delta, hc->tx_s, (unsigned)(hc->tx_x >> 6));
 }
 static u32 ccid3_hc_tx_idle_rtt(struct ccid3_hc_tx_sock *hc, ktime_t now)
@@ -211,16 +207,19 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data)
        ccid3_pr_debug("%s(%p, state=%s) - entry\n", dccp_role(sk), sk,
                       ccid3_tx_state_name(hc->tx_state));
+        /* Ignore and do not restart after leaving the established state */
+        if ((1 << sk->sk_state) & ~(DCCPF_OPEN | DCCPF_PARTOPEN))
+                goto out;
+        /* Reset feedback state to "no feedback received" */
        if (hc->tx_state == TFRC_SSTATE_FBACK)
                ccid3_hc_tx_set_state(sk, TFRC_SSTATE_NO_FBACK);
-        else if (hc->tx_state != TFRC_SSTATE_NO_FBACK)
-                goto out;
        /*
         * Determine new allowed sending rate X as per draft rfc3448bis-00, 4.4
+         * RTO is 0 if and only if no feedback has been received yet.
         */
-        if (hc->tx_t_rto == 0 ||        /* no feedback received yet */
+        if (hc->tx_t_rto == 0 || hc->tx_p == 0) {
-            hc->tx_p == 0) {
                /* halve send rate directly */
                hc->tx_x = max(hc->tx_x / 2,
@@ -256,7 +255,7 @@ static void ccid3_hc_tx_no_feedback_timer(unsigned long data)
         * Set new timeout for the nofeedback timer.
         * See comments in packet_recv() regarding the value of t_RTO.
         */
-        if (unlikely(hc->tx_t_rto == 0))        /* no feedback yet */
+        if (unlikely(hc->tx_t_rto == 0))        /* no feedback received yet */
                t_nfb = TFRC_INITIAL_TIMEOUT;
        else
                t_nfb = max(hc->tx_t_rto, 2 * hc->tx_t_ipi);
@@ -290,8 +289,7 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
        if (unlikely(skb->len == 0))
                return -EBADMSG;
-        switch (hc->tx_state) {
+        if (hc->tx_state == TFRC_SSTATE_NO_SENT) {
-        case TFRC_SSTATE_NO_SENT:
                sk_reset_timer(sk, &hc->tx_no_feedback_timer, (jiffies +
                               usecs_to_jiffies(TFRC_INITIAL_TIMEOUT)));
                hc->tx_last_win_count   = 0;
@@ -326,27 +324,22 @@ static int ccid3_hc_tx_send_packet(struct sock *sk, struct sk_buff *skb)
                ccid3_update_send_interval(hc);
                ccid3_hc_tx_set_state(sk, TFRC_SSTATE_NO_FBACK);
-                break;
-        case TFRC_SSTATE_NO_FBACK:
+        } else {
-        case TFRC_SSTATE_FBACK:
                delay = ktime_us_delta(hc->tx_t_nom, now);
                ccid3_pr_debug("delay=%ld\n", (long)delay);
                /*
-                 *      Scheduling of packet transmissions [RFC 3448, 4.6]
+                 *      Scheduling of packet transmissions (RFC 5348, 8.3)
                 *
                 * if (t_now > t_nom - delta)
                 *       // send the packet now
                 * else
                 *       // send the packet in (t_nom - t_now) milliseconds.
                 */
-                if (delay - (s64)hc->tx_delta >= 1000)
+                if (delay >= TFRC_T_DELTA)
-                        return (u32)delay / 1000L;
+                        return (u32)delay / USEC_PER_MSEC;
                ccid3_hc_tx_update_win_count(hc, now);
-                break;
-        case TFRC_SSTATE_TERM:
-                DCCP_BUG("%s(%p) - Illegal state TERM", dccp_role(sk), sk);
-                return -EINVAL;
        }
        /* prepare to send now (add options etc.) */
@@ -372,48 +365,34 @@ static void ccid3_hc_tx_packet_sent(struct sock *sk, int more,
 static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
 {
        struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk);
-        struct ccid3_options_received *opt_recv;
+        struct tfrc_tx_hist_entry *acked;
        ktime_t now;
        unsigned long t_nfb;
-        u32 pinv, r_sample;
+        u32 r_sample;
        /* we are only interested in ACKs */
        if (!(DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_ACK ||
              DCCP_SKB_CB(skb)->dccpd_type == DCCP_PKT_DATAACK))
                return;
-        /* ... and only in the established state */
-        if (hc->tx_state != TFRC_SSTATE_FBACK &&
-            hc->tx_state != TFRC_SSTATE_NO_FBACK)
-                return;
-        opt_recv = &hc->tx_options_received;
-        now = ktime_get_real();
-        /* Estimate RTT from history if ACK number is valid */
-        r_sample = tfrc_tx_hist_rtt(hc->tx_hist,
-                                    DCCP_SKB_CB(skb)->dccpd_ack_seq, now);
-        if (r_sample == 0) {
-                DCCP_WARN("%s(%p): %s with bogus ACK-%llu\n", dccp_role(sk), sk,
-                          dccp_packet_name(DCCP_SKB_CB(skb)->dccpd_type),
-                          (unsigned long long)DCCP_SKB_CB(skb)->dccpd_ack_seq);
-                return;
-        }
-        /* Update receive rate in units of 64 * bytes/second */
-        hc->tx_x_recv = opt_recv->ccid3or_receive_rate;
-        hc->tx_x_recv <<= 6;
-        /* Update loss event rate (which is scaled by 1e6) */
-        pinv = opt_recv->ccid3or_loss_event_rate;
-        if (pinv == ~0U || pinv == 0)          /* see RFC 4342, 8.5   */
-                hc->tx_p = 0;
-        else                                   /* can not exceed 100% */
-                hc->tx_p = scaled_div(1, pinv);
        /*
-         * Validate new RTT sample and update moving average
+         * Locate the acknowledged packet in the TX history.
+         *
+         * Returning "entry not found" here can for instance happen when
+         *  - the host has not sent out anything (e.g. a passive server),
+         *  - the Ack is outdated (packet with higher Ack number was received),
+         *  - it is a bogus Ack (for a packet not sent on this connection).
         */
-        r_sample = dccp_sample_rtt(sk, r_sample);
+        acked = tfrc_tx_hist_find_entry(hc->tx_hist, dccp_hdr_ack_seq(skb));
+        if (acked == NULL)
+                return;
+        /* For the sake of RTT sampling, ignore/remove all older entries */
+        tfrc_tx_hist_purge(&acked->next);
+        /* Update the moving average for the RTT estimate (RFC 3448, 4.3) */
+        now       = ktime_get_real();
+        r_sample  = dccp_sample_rtt(sk, ktime_us_delta(now, acked->stamp));
        hc->tx_rtt = tfrc_ewma(hc->tx_rtt, r_sample, 9);
        /*
         * Update allowed sending rate X as per draft rfc3448bis-00, 4.2/3
         */
@@ -461,13 +440,12 @@ done_computing_x:
        sk->sk_write_space(sk);
        /*
-         * Update timeout interval for the nofeedback timer.
+         * Update timeout interval for the nofeedback timer. In order to control
-         * We use a configuration option to increase the lower bound.
+         * rate halving on networks with very low RTTs (<= 1 ms), use per-route
-         * This can help avoid triggering the nofeedback timer too
+         * tunable RTAX_RTO_MIN value as the lower bound.
-         * often ('spinning') on LANs with small RTTs.
         */
-        hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt, (CONFIG_IP_DCCP_CCID3_RTO *
+        hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt,
-                                                       (USEC_PER_SEC / 1000)));
+                                  USEC_PER_SEC/HZ * tcp_rto_min(sk));
        /*
         * Schedule no feedback timer to expire in
         * max(t_RTO, 2 * s/X)  =  max(t_RTO, 2 * t_ipi)
@@ -482,66 +460,41 @@ done_computing_x:
                           jiffies + usecs_to_jiffies(t_nfb));
 }
-static int ccid3_hc_tx_parse_options(struct sock *sk, unsigned char option,
+static int ccid3_hc_tx_parse_options(struct sock *sk, u8 packet_type,
-                                     unsigned char len, u16 idx,
+                                     u8 option, u8 *optval, u8 optlen)
-                                     unsigned char *value)
 {
-        int rc = 0;
-        const struct dccp_sock *dp = dccp_sk(sk);
        struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk);
-        struct ccid3_options_received *opt_recv;
        __be32 opt_val;
-        opt_recv = &hc->tx_options_received;
-        if (opt_recv->ccid3or_seqno != dp->dccps_gsr) {
-                opt_recv->ccid3or_seqno              = dp->dccps_gsr;
-                opt_recv->ccid3or_loss_event_rate    = ~0;
-                opt_recv->ccid3or_loss_intervals_idx = 0;
-                opt_recv->ccid3or_loss_intervals_len = 0;
-                opt_recv->ccid3or_receive_rate       = 0;
-        }
        switch (option) {
+        case TFRC_OPT_RECEIVE_RATE:
        case TFRC_OPT_LOSS_EVENT_RATE:
-                if (unlikely(len != 4)) {
+                /* Must be ignored on Data packets, cf. RFC 4342 8.3 and 8.5 */
-                        DCCP_WARN("%s(%p), invalid len %d "
+                if (packet_type == DCCP_PKT_DATA)
-                                  "for TFRC_OPT_LOSS_EVENT_RATE\n",
+                        break;
-                                  dccp_role(sk), sk, len);
+                if (unlikely(optlen != 4)) {
-                        rc = -EINVAL;
+                        DCCP_WARN("%s(%p), invalid len %d for %u\n",
-                } else {
+                                  dccp_role(sk), sk, optlen, option);
-                        opt_val = get_unaligned((__be32 *)value);
+                        return -EINVAL;
-                        opt_recv->ccid3or_loss_event_rate = ntohl(opt_val);
-                        ccid3_pr_debug("%s(%p), LOSS_EVENT_RATE=%u\n",
-                                       dccp_role(sk), sk,
-                                       opt_recv->ccid3or_loss_event_rate);
                }
-                break;
+                opt_val = ntohl(get_unaligned((__be32 *)optval));
-        case TFRC_OPT_LOSS_INTERVALS:
-                opt_recv->ccid3or_loss_intervals_idx = idx;
+                if (option == TFRC_OPT_RECEIVE_RATE) {
-                opt_recv->ccid3or_loss_intervals_len = len;
+                        /* Receive Rate is kept in units of 64 bytes/second */
-                ccid3_pr_debug("%s(%p), LOSS_INTERVALS=(%u, %u)\n",
+                        hc->tx_x_recv = opt_val;
-                               dccp_role(sk), sk,
+                        hc->tx_x_recv <<= 6;
-                               opt_recv->ccid3or_loss_intervals_idx,
-                               opt_recv->ccid3or_loss_intervals_len);
-                break;
-        case TFRC_OPT_RECEIVE_RATE:
-                if (unlikely(len != 4)) {
-                        DCCP_WARN("%s(%p), invalid len %d "
-                                  "for TFRC_OPT_RECEIVE_RATE\n",
-                                  dccp_role(sk), sk, len);
-                        rc = -EINVAL;
-                } else {
-                        opt_val = get_unaligned((__be32 *)value);
-                        opt_recv->ccid3or_receive_rate = ntohl(opt_val);
                        ccid3_pr_debug("%s(%p), RECEIVE_RATE=%u\n",
-                                       dccp_role(sk), sk,
+                                       dccp_role(sk), sk, opt_val);
-                                       opt_recv->ccid3or_receive_rate);
+                } else {
+                        /* Update the fixpoint Loss Event Rate fraction */
+                        hc->tx_p = tfrc_invert_loss_event_rate(opt_val);
+                        ccid3_pr_debug("%s(%p), LOSS_EVENT_RATE=%u\n",
+                                       dccp_role(sk), sk, opt_val);
                }
-                break;
        }
+        return 0;
-        return rc;
 }
 static int ccid3_hc_tx_init(struct ccid *ccid, struct sock *sk)
@@ -559,42 +512,36 @@ static void ccid3_hc_tx_exit(struct sock *sk)
 {
        struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk);
-        ccid3_hc_tx_set_state(sk, TFRC_SSTATE_TERM);
        sk_stop_timer(sk, &hc->tx_no_feedback_timer);
        tfrc_tx_hist_purge(&hc->tx_hist);
 }
 static void ccid3_hc_tx_get_info(struct sock *sk, struct tcp_info *info)
 {
-        struct ccid3_hc_tx_sock *hc;
+        info->tcpi_rto = ccid3_hc_tx_sk(sk)->tx_t_rto;
+        info->tcpi_rtt = ccid3_hc_tx_sk(sk)->tx_rtt;
-        /* Listen socks doesn't have a private CCID block */
-        if (sk->sk_state == DCCP_LISTEN)
-                return;
-        hc = ccid3_hc_tx_sk(sk);
-        info->tcpi_rto = hc->tx_t_rto;
-        info->tcpi_rtt = hc->tx_rtt;
 }
 static int ccid3_hc_tx_getsockopt(struct sock *sk, const int optname, int len,
                                  u32 __user *optval, int __user *optlen)
 {
-        const struct ccid3_hc_tx_sock *hc;
+        const struct ccid3_hc_tx_sock *hc = ccid3_hc_tx_sk(sk);
+        struct tfrc_tx_info tfrc;
        const void *val;
-        /* Listen socks doesn't have a private CCID block */
-        if (sk->sk_state == DCCP_LISTEN)
-                return -EINVAL;
-        hc = ccid3_hc_tx_sk(sk);
        switch (optname) {
        case DCCP_SOCKOPT_CCID_TX_INFO:
-                if (len < sizeof(hc->tx_tfrc))
+                if (len < sizeof(tfrc))
                        return -EINVAL;
-                len = sizeof(hc->tx_tfrc);
+                tfrc.tfrctx_x      = hc->tx_x;
-                val = &hc->tx_tfrc;
+                tfrc.tfrctx_x_recv = hc->tx_x_recv;
+                tfrc.tfrctx_x_calc = hc->tx_x_calc;
+                tfrc.tfrctx_rtt    = hc->tx_rtt;
+                tfrc.tfrctx_p      = hc->tx_p;
+                tfrc.tfrctx_rto    = hc->tx_t_rto;
+                tfrc.tfrctx_ipi    = hc->tx_t_ipi;
+                len = sizeof(tfrc);
+                val = &tfrc;
                break;
        default:
                return -ENOPROTOOPT;
@@ -624,7 +571,6 @@ static const char *ccid3_rx_state_name(enum ccid3_hc_rx_states state)
        static const char *const ccid3_rx_state_names[] = {
        [TFRC_RSTATE_NO_DATA] = "NO_DATA",
        [TFRC_RSTATE_DATA]    = "DATA",
-        [TFRC_RSTATE_TERM]    = "TERM",
        };
        return ccid3_rx_state_names[state];
@@ -650,14 +596,9 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk,
 {
        struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk);
        struct dccp_sock *dp = dccp_sk(sk);
-        ktime_t now;
+        ktime_t now = ktime_get_real();
        s64 delta = 0;
-        if (unlikely(hc->rx_state == TFRC_RSTATE_TERM))
-                return;
-        now = ktime_get_real();
        switch (fbtype) {
        case CCID3_FBACK_INITIAL:
                hc->rx_x_recv = 0;
@@ -701,14 +642,12 @@ static void ccid3_hc_rx_send_feedback(struct sock *sk,
 static int ccid3_hc_rx_insert_options(struct sock *sk, struct sk_buff *skb)
 {
-        const struct ccid3_hc_rx_sock *hc;
+        const struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk);
        __be32 x_recv, pinv;
        if (!(sk->sk_state == DCCP_OPEN || sk->sk_state == DCCP_PARTOPEN))
                return 0;
-        hc = ccid3_hc_rx_sk(sk);
        if (dccp_packet_without_ack(skb))
                return 0;
@@ -749,10 +688,11 @@ static u32 ccid3_first_li(struct sock *sk)
        x_recv = scaled_div32(hc->rx_bytes_recv, delta);
        if (x_recv == 0) {              /* would also trigger divide-by-zero */
                DCCP_WARN("X_recv==0\n");
-                if ((x_recv = hc->rx_x_recv) == 0) {
+                if (hc->rx_x_recv == 0) {
                        DCCP_BUG("stored value of X_recv is zero");
                        return ~0U;
                }
+                x_recv = hc->rx_x_recv;
        }
        fval = scaled_div(hc->rx_s, hc->rx_rtt);
@@ -862,46 +802,31 @@ static void ccid3_hc_rx_exit(struct sock *sk)
 {
        struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk);
-        ccid3_hc_rx_set_state(sk, TFRC_RSTATE_TERM);
        tfrc_rx_hist_purge(&hc->rx_hist);
        tfrc_lh_cleanup(&hc->rx_li_hist);
 }
 static void ccid3_hc_rx_get_info(struct sock *sk, struct tcp_info *info)
 {
-        const struct ccid3_hc_rx_sock *hc;
+        info->tcpi_ca_state = ccid3_hc_rx_sk(sk)->rx_state;
-        /* Listen socks doesn't have a private CCID block */
-        if (sk->sk_state == DCCP_LISTEN)
-                return;
-        hc = ccid3_hc_rx_sk(sk);
-        info->tcpi_ca_state = hc->rx_state;
        info->tcpi_options  |= TCPI_OPT_TIMESTAMPS;
-        info->tcpi_rcv_rtt  = hc->rx_rtt;
+        info->tcpi_rcv_rtt  = ccid3_hc_rx_sk(sk)->rx_rtt;
 }
 static int ccid3_hc_rx_getsockopt(struct sock *sk, const int optname, int len,
                                  u32 __user *optval, int __user *optlen)
 {
-        const struct ccid3_hc_rx_sock *hc;
+        const struct ccid3_hc_rx_sock *hc = ccid3_hc_rx_sk(sk);
        struct tfrc_rx_info rx_info;
        const void *val;
-        /* Listen socks doesn't have a private CCID block */
-        if (sk->sk_state == DCCP_LISTEN)
-                return -EINVAL;
-        hc = ccid3_hc_rx_sk(sk);
        switch (optname) {
        case DCCP_SOCKOPT_CCID_RX_INFO:
                if (len < sizeof(rx_info))
                        return -EINVAL;
                rx_info.tfrcrx_x_recv = hc->rx_x_recv;
                rx_info.tfrcrx_rtt    = hc->rx_rtt;
-                rx_info.tfrcrx_p      = hc->rx_pinv == 0 ? ~0U :
+                rx_info.tfrcrx_p      = tfrc_invert_loss_event_rate(hc->rx_pinv);
-                                           scaled_div(1, hc->rx_pinv);
                len = sizeof(rx_info);
                val = &rx_info;
                break;
diff --git a/net/dccp/ccids/ccid3.h b/net/dccp/ccids/ccid3.h
index 032635776653..1a9933c29672 100644
--- a/net/dccp/ccids/ccid3.h
+++ b/net/dccp/ccids/ccid3.h
@@ -42,35 +42,36 @@
 #include "lib/tfrc.h"
 #include "../ccid.h"
-/* Two seconds as per RFC 3448 4.2 */
+/* Two seconds as per RFC 5348, 4.2 */
 #define TFRC_INITIAL_TIMEOUT       (2 * USEC_PER_SEC)
-/* In usecs - half the scheduling granularity as per RFC3448 4.6 */
-#define TFRC_OPSYS_HALF_TIME_GRAN  (USEC_PER_SEC / (2 * HZ))
 /* Parameter t_mbi from [RFC 3448, 4.3]: backoff interval in seconds */
 #define TFRC_T_MBI                 64
+/*
+ * The t_delta parameter (RFC 5348, 8.3): delays of less than %USEC_PER_MSEC are
+ * rounded down to 0, since sk_reset_timer() here uses millisecond granularity.
+ * Hence we can use a constant t_delta = %USEC_PER_MSEC when HZ >= 500. A coarse
+ * resolution of HZ < 500 means that the error is below one timer tick (t_gran)
+ * when using the constant t_delta  =  t_gran / 2  =  %USEC_PER_SEC / (2 * HZ).
+ */
+#if (HZ >= 500)
+# define TFRC_T_DELTA              USEC_PER_MSEC
+#else
+# define TFRC_T_DELTA              (USEC_PER_SEC / (2 * HZ))
+#endif
 enum ccid3_options {
        TFRC_OPT_LOSS_EVENT_RATE = 192,
        TFRC_OPT_LOSS_INTERVALS  = 193,
        TFRC_OPT_RECEIVE_RATE    = 194,
 };
-struct ccid3_options_received {
-        u64 ccid3or_seqno:48,
-            ccid3or_loss_intervals_idx:16;
-        u16 ccid3or_loss_intervals_len;
-        u32 ccid3or_loss_event_rate;
-        u32 ccid3or_receive_rate;
-};
 /* TFRC sender states */
 enum ccid3_hc_tx_states {
        TFRC_SSTATE_NO_SENT = 1,
        TFRC_SSTATE_NO_FBACK,
        TFRC_SSTATE_FBACK,
-        TFRC_SSTATE_TERM,
 };
 /**
@@ -90,19 +91,16 @@ enum ccid3_hc_tx_states {
 * @tx_no_feedback_timer: Handle to no feedback timer
 * @tx_t_ld:              Time last doubled during slow start
 * @tx_t_nom:             Nominal send time of next packet
- * @tx_delta:             Send timer delta (RFC 3448, 4.6) in usecs
 * @tx_hist:              Packet history
- * @tx_options_received:  Parsed set of retrieved options
 */
 struct ccid3_hc_tx_sock {
-        struct tfrc_tx_info             tx_tfrc;
+        u64                             tx_x;
-#define tx_x                            tx_tfrc.tfrctx_x
+        u64                             tx_x_recv;
-#define tx_x_recv                       tx_tfrc.tfrctx_x_recv
+        u32                             tx_x_calc;
-#define tx_x_calc                       tx_tfrc.tfrctx_x_calc
+        u32                             tx_rtt;
-#define tx_rtt                          tx_tfrc.tfrctx_rtt
+        u32                             tx_p;
-#define tx_p                            tx_tfrc.tfrctx_p
+        u32                             tx_t_rto;
-#define tx_t_rto                        tx_tfrc.tfrctx_rto
+        u32                             tx_t_ipi;
-#define tx_t_ipi                        tx_tfrc.tfrctx_ipi
        u16                             tx_s;
        enum ccid3_hc_tx_states         tx_state:8;
        u8                              tx_last_win_count;
@@ -110,9 +108,7 @@ struct ccid3_hc_tx_sock {
        struct timer_list               tx_no_feedback_timer;
        ktime_t                         tx_t_ld;
        ktime_t                         tx_t_nom;
-        u32                             tx_delta;
        struct tfrc_tx_hist_entry       *tx_hist;
-        struct ccid3_options_received   tx_options_received;
 };
 static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk)
@@ -126,21 +122,16 @@ static inline struct ccid3_hc_tx_sock *ccid3_hc_tx_sk(const struct sock *sk)
 enum ccid3_hc_rx_states {
        TFRC_RSTATE_NO_DATA = 1,
        TFRC_RSTATE_DATA,
-        TFRC_RSTATE_TERM    = 127,
 };
 /**
 * struct ccid3_hc_rx_sock - CCID3 receiver half-connection socket
- * @rx_x_recv:               Receiver estimate of send rate (RFC 3448 4.3)
- * @rx_rtt:                  Receiver estimate of rtt (non-standard)
- * @rx_p:                    Current loss event rate (RFC 3448 5.4)
 * @rx_last_counter:         Tracks window counter (RFC 4342, 8.1)
 * @rx_state:                Receiver state, one of %ccid3_hc_rx_states
 * @rx_bytes_recv:           Total sum of DCCP payload bytes
 * @rx_x_recv:               Receiver estimate of send rate (RFC 3448, sec. 4.3)
 * @rx_rtt:                  Receiver estimate of RTT
 * @rx_tstamp_last_feedback: Time at which last feedback was sent
- * @rx_tstamp_last_ack:      Time at which last feedback was sent
 * @rx_hist:                 Packet history (loss detection + RTT sampling)
 * @rx_li_hist:              Loss Interval database
 * @rx_s:                    Received packet size in bytes
diff --git a/net/dccp/ccids/lib/loss_interval.c b/net/dccp/ccids/lib/loss_interval.c
index 8fc3cbf79071..497723c4d4bb 100644
--- a/net/dccp/ccids/lib/loss_interval.c
+++ b/net/dccp/ccids/lib/loss_interval.c
@@ -116,7 +116,7 @@ u8 tfrc_lh_update_i_mean(struct tfrc_loss_hist *lh, struct sk_buff *skb)
        cur->li_length = len;
        tfrc_lh_calc_i_mean(lh);
-        return (lh->i_mean < old_i_mean);
+        return lh->i_mean < old_i_mean;
 }
 /* Determine if `new_loss' does begin a new loss interval [RFC 4342, 10.2] */
diff --git a/net/dccp/ccids/lib/packet_history.c b/net/dccp/ccids/lib/packet_history.c
index 3a4f414e94a0..de8fe294bf0b 100644
--- a/net/dccp/ccids/lib/packet_history.c
+++ b/net/dccp/ccids/lib/packet_history.c
@@ -38,18 +38,6 @@
 #include "packet_history.h"
 #include "../../dccp.h"
-/**
- *  tfrc_tx_hist_entry  -  Simple singly-linked TX history list
- *  @next:  next oldest entry (LIFO order)
- *  @seqno: sequence number of this entry
- *  @stamp: send time of packet with sequence number @seqno
- */
-struct tfrc_tx_hist_entry {
-        struct tfrc_tx_hist_entry *next;
-        u64                       seqno;
-        ktime_t                   stamp;
-};
 /*
 * Transmitter History Routines
 */
@@ -71,15 +59,6 @@ void tfrc_tx_packet_history_exit(void)
        }
 }
-static struct tfrc_tx_hist_entry *
-        tfrc_tx_hist_find_entry(struct tfrc_tx_hist_entry *head, u64 seqno)
-{
-        while (head != NULL && head->seqno != seqno)
-                head = head->next;
-        return head;
-}
 int tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno)
 {
        struct tfrc_tx_hist_entry *entry = kmem_cache_alloc(tfrc_tx_hist_slab, gfp_any());
@@ -107,24 +86,6 @@ void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp)
        *headp = NULL;
 }
-u32 tfrc_tx_hist_rtt(struct tfrc_tx_hist_entry *head, const u64 seqno,
-                     const ktime_t now)
-{
-        u32 rtt = 0;
-        struct tfrc_tx_hist_entry *packet = tfrc_tx_hist_find_entry(head, seqno);
-        if (packet != NULL) {
-                rtt = ktime_us_delta(now, packet->stamp);
-                /*
-                 * Garbage-collect older (irrelevant) entries:
-                 */
-                tfrc_tx_hist_purge(&packet->next);
-        }
-        return rtt;
-}
 /*
 *      Receiver History Routines
 */
diff --git a/net/dccp/ccids/lib/packet_history.h b/net/dccp/ccids/lib/packet_history.h
index 7df6c5299999..7ee4a9d9d335 100644
--- a/net/dccp/ccids/lib/packet_history.h
+++ b/net/dccp/ccids/lib/packet_history.h
@@ -40,12 +40,28 @@
 #include <linux/slab.h>
 #include "tfrc.h"
-struct tfrc_tx_hist_entry;
+/**
+ *  tfrc_tx_hist_entry  -  Simple singly-linked TX history list
+ *  @next:  next oldest entry (LIFO order)
+ *  @seqno: sequence number of this entry
+ *  @stamp: send time of packet with sequence number @seqno
+ */
+struct tfrc_tx_hist_entry {
+        struct tfrc_tx_hist_entry *next;
+        u64                       seqno;
+        ktime_t                   stamp;
+};
+static inline struct tfrc_tx_hist_entry *
+        tfrc_tx_hist_find_entry(struct tfrc_tx_hist_entry *head, u64 seqno)
+{
+        while (head != NULL && head->seqno != seqno)
+                head = head->next;
+        return head;
+}
 extern int  tfrc_tx_hist_add(struct tfrc_tx_hist_entry **headp, u64 seqno);
 extern void tfrc_tx_hist_purge(struct tfrc_tx_hist_entry **headp);
-extern u32  tfrc_tx_hist_rtt(struct tfrc_tx_hist_entry *head,
-                             const u64 seqno, const ktime_t now);
 /* Subtraction a-b modulo-16, respects circular wrap-around */
 #define SUB16(a, b) (((a) + 16 - (b)) & 0xF)
diff --git a/net/dccp/ccids/lib/tfrc.h b/net/dccp/ccids/lib/tfrc.h
index 01bb48e96c2e..f8ee3f549770 100644
--- a/net/dccp/ccids/lib/tfrc.h
+++ b/net/dccp/ccids/lib/tfrc.h
@@ -57,6 +57,7 @@ static inline u32 tfrc_ewma(const u32 avg, const u32 newval, const u8 weight)
 extern u32  tfrc_calc_x(u16 s, u32 R, u32 p);
 extern u32  tfrc_calc_x_reverse_lookup(u32 fvalue);
+extern u32  tfrc_invert_loss_event_rate(u32 loss_event_rate);
 extern int  tfrc_tx_packet_history_init(void);
 extern void tfrc_tx_packet_history_exit(void);
diff --git a/net/dccp/ccids/lib/tfrc_equation.c b/net/dccp/ccids/lib/tfrc_equation.c
index 22ca1cf0eb55..a052a4377e26 100644
--- a/net/dccp/ccids/lib/tfrc_equation.c
+++ b/net/dccp/ccids/lib/tfrc_equation.c
@@ -687,3 +687,17 @@ u32 tfrc_calc_x_reverse_lookup(u32 fvalue)
        index = tfrc_binsearch(fvalue, 0);
        return (index + 1) * 1000000 / TFRC_CALC_X_ARRSIZE;
 }
+/**
+ * tfrc_invert_loss_event_rate  -  Compute p so that 10^6 corresponds to 100%
+ * When @loss_event_rate is large, there is a chance that p is truncated to 0.
+ * To avoid re-entering slow-start in that case, we set p = TFRC_SMALLEST_P > 0.
+ */
+u32 tfrc_invert_loss_event_rate(u32 loss_event_rate)
+{
+        if (loss_event_rate == UINT_MAX)                /* see RFC 4342, 8.5 */
+                return 0;
+        if (unlikely(loss_event_rate == 0))             /* map 1/0 into 100% */
+                return 1000000;
+        return max_t(u32, scaled_div(1, loss_event_rate), TFRC_SMALLEST_P);
+}
diff --git a/net/dccp/options.c b/net/dccp/options.c
index bfda087bd90d..92718511eac5 100644
--- a/net/dccp/options.c
+++ b/net/dccp/options.c
@@ -96,18 +96,11 @@ int dccp_parse_options(struct sock *sk, struct dccp_request_sock *dreq,
                }
                /*
-                 * CCID-Specific Options (from RFC 4340, sec. 10.3):
-                 *
-                 * Option numbers 128 through 191 are for options sent from the
-                 * HC-Sender to the HC-Receiver; option numbers 192 through 255
-                 * are for options sent from the HC-Receiver to the HC-Sender.
-                 *
                 * CCID-specific options are ignored during connection setup, as
                 * negotiation may still be in progress (see RFC 4340, 10.3).
                 * The same applies to Ack Vectors, as these depend on the CCID.
-                 *
                 */
-                if (dreq != NULL && (opt >= 128 ||
+                if (dreq != NULL && (opt >= DCCPO_MIN_RX_CCID_SPECIFIC ||
                    opt == DCCPO_ACK_VECTOR_0 || opt == DCCPO_ACK_VECTOR_1))
                        goto ignore_option;
@@ -226,23 +219,15 @@ int dccp_parse_options(struct sock *sk, struct dccp_request_sock *dreq,
                        dccp_pr_debug("%s rx opt: ELAPSED_TIME=%d\n",
                                      dccp_role(sk), elapsed_time);
                        break;
-                case 128 ... 191: {
+                case DCCPO_MIN_RX_CCID_SPECIFIC ... DCCPO_MAX_RX_CCID_SPECIFIC:
-                        const u16 idx = value - options;
                        if (ccid_hc_rx_parse_options(dp->dccps_hc_rx_ccid, sk,
-                                                     opt, len, idx,
+                                                     pkt_type, opt, value, len))
-                                                     value) != 0)
                                goto out_invalid_option;
-                }
                        break;
-                case 192 ... 255: {
+                case DCCPO_MIN_TX_CCID_SPECIFIC ... DCCPO_MAX_TX_CCID_SPECIFIC:
-                        const u16 idx = value - options;
                        if (ccid_hc_tx_parse_options(dp->dccps_hc_tx_ccid, sk,
-                                                     opt, len, idx,
+                                                     pkt_type, opt, value, len))
-                                                     value) != 0)
                                goto out_invalid_option;
-                }
                        break;
                default:
                        DCCP_CRIT("DCCP(%p): option %d(len=%d) not "
diff --git a/net/decnet/dn_nsp_out.c b/net/decnet/dn_nsp_out.c
index baeb1eaf011b..2ef115277bea 100644
--- a/net/decnet/dn_nsp_out.c
+++ b/net/decnet/dn_nsp_out.c
@@ -693,22 +693,22 @@ void dn_nsp_send_conninit(struct sock *sk, unsigned char msgflg)
        aux = scp->accessdata.acc_userl;
        *skb_put(skb, 1) = aux;
        if (aux > 0)
-        memcpy(skb_put(skb, aux), scp->accessdata.acc_user, aux);
+                memcpy(skb_put(skb, aux), scp->accessdata.acc_user, aux);
        aux = scp->accessdata.acc_passl;
        *skb_put(skb, 1) = aux;
        if (aux > 0)
-        memcpy(skb_put(skb, aux), scp->accessdata.acc_pass, aux);
+                memcpy(skb_put(skb, aux), scp->accessdata.acc_pass, aux);
        aux = scp->accessdata.acc_accl;
        *skb_put(skb, 1) = aux;
        if (aux > 0)
-        memcpy(skb_put(skb, aux), scp->accessdata.acc_acc, aux);
+                memcpy(skb_put(skb, aux), scp->accessdata.acc_acc, aux);
        aux = (__u8)le16_to_cpu(scp->conndata_out.opt_optl);
        *skb_put(skb, 1) = aux;
        if (aux > 0)
-        memcpy(skb_put(skb,aux), scp->conndata_out.opt_data, aux);
+                memcpy(skb_put(skb, aux), scp->conndata_out.opt_data, aux);
        scp->persist = dn_nsp_persist(sk);
        scp->persist_fxn = dn_nsp_retrans_conninit;
diff --git a/net/econet/af_econet.c b/net/econet/af_econet.c
index dc54bd0d083b..f8c1ae4b41f0 100644
--- a/net/econet/af_econet.c
+++ b/net/econet/af_econet.c
@@ -392,7 +392,7 @@ static int econet_sendmsg(struct kiocb *iocb, struct socket *sock,
                dev_queue_xmit(skb);
                dev_put(dev);
                mutex_unlock(&econet_mutex);
-                return(len);
+                return len;
        out_free:
                kfree_skb(skb);
@@ -637,7 +637,7 @@ static int econet_create(struct net *net, struct socket *sock, int protocol,
        eo->num = protocol;
        econet_insert_socket(&econet_sklist, sk);
-        return(0);
+        return 0;
 out:
        return err;
 }
@@ -1009,7 +1009,6 @@ static int __init aun_udp_initialise(void)
        struct sockaddr_in sin;
        skb_queue_head_init(&aun_queue);
-        spin_lock_init(&aun_queue_lock);
        setup_timer(&ab_cleanup_timer, ab_cleanup, 0);
        ab_cleanup_timer.expires = jiffies + (HZ*2);
        add_timer(&ab_cleanup_timer);
@@ -1167,7 +1166,6 @@ static int __init econet_proto_init(void)
                goto out;
        sock_register(&econet_family_ops);
 #ifdef CONFIG_ECONET_AUNUDP
-        spin_lock_init(&aun_queue_lock);
        aun_udp_initialise();
 #endif
 #ifdef CONFIG_ECONET_NATIVE
diff --git a/net/ethernet/eth.c b/net/ethernet/eth.c
index 215c83986a9d..f00ef2f1d814 100644
--- a/net/ethernet/eth.c
+++ b/net/ethernet/eth.c
@@ -367,7 +367,7 @@ struct net_device *alloc_etherdev_mq(int sizeof_priv, unsigned int queue_count)
 EXPORT_SYMBOL(alloc_etherdev_mq);
 static size_t _format_mac_addr(char *buf, int buflen,
-                                const unsigned char *addr, int len)
+                               const unsigned char *addr, int len)
 {
        int i;
        char *cp = buf;
@@ -376,7 +376,7 @@ static size_t _format_mac_addr(char *buf, int buflen,
                cp += scnprintf(cp, buflen - (cp - buf), "%02x", addr[i]);
                if (i == len - 1)
                        break;
-                cp += strlcpy(cp, ":", buflen - (cp - buf));
+                cp += scnprintf(cp, buflen - (cp - buf), ":");
        }
        return cp - buf;
 }
@@ -386,7 +386,7 @@ ssize_t sysfs_format_mac(char *buf, const unsigned char *addr, int len)
        size_t l;
        l = _format_mac_addr(buf, PAGE_SIZE, addr, len);
-        l += strlcpy(buf + l, "\n", PAGE_SIZE - l);
+        l += scnprintf(buf + l, PAGE_SIZE - l, "\n");
-        return ((ssize_t) l);
+        return (ssize_t)l;
 }
 EXPORT_SYMBOL(sysfs_format_mac);
diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig
index 7cd7760144f7..e848e6c062cd 100644
--- a/net/ipv4/Kconfig
+++ b/net/ipv4/Kconfig
@@ -215,9 +215,15 @@ config NET_IPIP
          be inserted in and removed from the running kernel whenever you
          want). Most people won't need this and can say N.
+config NET_IPGRE_DEMUX
+        tristate "IP: GRE demultiplexer"
+        help
+         This is helper module to demultiplex GRE packets on GRE version field criteria.
+         Required by ip_gre and pptp modules.
 config NET_IPGRE
        tristate "IP: GRE tunnels over IP"
-        depends on IPV6 || IPV6=n
+        depends on (IPV6 || IPV6=n) && NET_IPGRE_DEMUX
        help
          Tunneling means encapsulating data of one protocol type within
          another protocol and sending it over a channel that understands the
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index 80ff87ce43aa..4978d22f9a75 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -20,6 +20,7 @@ obj-$(CONFIG_PROC_FS) += proc.o
 obj-$(CONFIG_IP_MULTIPLE_TABLES) += fib_rules.o
 obj-$(CONFIG_IP_MROUTE) += ipmr.o
 obj-$(CONFIG_NET_IPIP) += ipip.o
+obj-$(CONFIG_NET_IPGRE_DEMUX) += gre.o
 obj-$(CONFIG_NET_IPGRE) += ip_gre.o
 obj-$(CONFIG_SYN_COOKIES) += syncookies.o
 obj-$(CONFIG_INET_AH) += ah4.o
diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 6a1100c25a9f..f581f77d1097 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -227,18 +227,16 @@ EXPORT_SYMBOL(inet_ehash_secret);
 /*
 * inet_ehash_secret must be set exactly once
- * Instead of using a dedicated spinlock, we (ab)use inetsw_lock
 */
 void build_ehash_secret(void)
 {
        u32 rnd;
        do {
                get_random_bytes(&rnd, sizeof(rnd));
        } while (rnd == 0);
-        spin_lock_bh(&inetsw_lock);
-        if (!inet_ehash_secret)
+        cmpxchg(&inet_ehash_secret, 0, rnd);
-                inet_ehash_secret = rnd;
-        spin_unlock_bh(&inetsw_lock);
 }
 EXPORT_SYMBOL(build_ehash_secret);
diff --git a/net/ipv4/arp.c b/net/ipv4/arp.c
index 96c1955b3e2f..d9031ad67826 100644
--- a/net/ipv4/arp.c
+++ b/net/ipv4/arp.c
@@ -55,7 +55,7 @@
 *              Stuart Cheshire :       Metricom and grat arp fixes
 *                                      *** FOR 2.1 clean this up ***
 *              Lawrence V. Stefani: (08/12/96) Added FDDI support.
- *              Alan Cox        :       Took the AP1000 nasty FDDI hack and
+ *              Alan Cox        :       Took the AP1000 nasty FDDI hack and
 *                                      folded into the mainstream FDDI code.
 *                                      Ack spit, Linus how did you allow that
 *                                      one in...
@@ -120,7 +120,7 @@ EXPORT_SYMBOL(clip_tbl_hook);
 #endif
 #include <asm/system.h>
-#include <asm/uaccess.h>
+#include <linux/uaccess.h>
 #include <linux/netfilter_arp.h>
@@ -161,7 +161,7 @@ static const struct neigh_ops arp_direct_ops = {
        .queue_xmit =           dev_queue_xmit,
 };
-const struct neigh_ops arp_broken_ops = {
+static const struct neigh_ops arp_broken_ops = {
        .family =               AF_INET,
        .solicit =              arp_solicit,
        .error_report =         arp_error_report,
@@ -170,35 +170,34 @@ const struct neigh_ops arp_broken_ops = {
        .hh_output =            dev_queue_xmit,
        .queue_xmit =           dev_queue_xmit,
 };
-EXPORT_SYMBOL(arp_broken_ops);
 struct neigh_table arp_tbl = {
-        .family =       AF_INET,
+        .family         = AF_INET,
-        .entry_size =   sizeof(struct neighbour) + 4,
+        .entry_size     = sizeof(struct neighbour) + 4,
-        .key_len =      4,
+        .key_len        = 4,
-        .hash =         arp_hash,
+        .hash           = arp_hash,
-        .constructor =  arp_constructor,
+        .constructor    = arp_constructor,
-        .proxy_redo =   parp_redo,
+        .proxy_redo     = parp_redo,
-        .id =           "arp_cache",
+        .id             = "arp_cache",
-        .parms = {
+        .parms          = {
-                .tbl =                  &arp_tbl,
+                .tbl                    = &arp_tbl,
-                .base_reachable_time =  30 * HZ,
+                .base_reachable_time    = 30 * HZ,
-                .retrans_time = 1 * HZ,
+                .retrans_time           = 1 * HZ,
-                .gc_staletime = 60 * HZ,
+                .gc_staletime           = 60 * HZ,
-                .reachable_time =               30 * HZ,
+                .reachable_time         = 30 * HZ,
-                .delay_probe_time =     5 * HZ,
+                .delay_probe_time       = 5 * HZ,
-                .queue_len =            3,
+                .queue_len              = 3,
-                .ucast_probes = 3,
+                .ucast_probes           = 3,
-                .mcast_probes = 3,
+                .mcast_probes           = 3,
-                .anycast_delay =        1 * HZ,
+                .anycast_delay          = 1 * HZ,
-                .proxy_delay =          (8 * HZ) / 10,
+                .proxy_delay            = (8 * HZ) / 10,
-                .proxy_qlen =           64,
+                .proxy_qlen             = 64,
-                .locktime =             1 * HZ,
+                .locktime               = 1 * HZ,
        },
-        .gc_interval =  30 * HZ,
+        .gc_interval    = 30 * HZ,
-        .gc_thresh1 =   128,
+        .gc_thresh1     = 128,
-        .gc_thresh2 =   512,
+        .gc_thresh2     = 512,
-        .gc_thresh3 =   1024,
+        .gc_thresh3     = 1024,
 };
 EXPORT_SYMBOL(arp_tbl);
@@ -233,7 +232,7 @@ static u32 arp_hash(const void *pkey, const struct net_device *dev)
 static int arp_constructor(struct neighbour *neigh)
 {
-        __be32 addr = *(__be32*)neigh->primary_key;
+        __be32 addr = *(__be32 *)neigh->primary_key;
        struct net_device *dev = neigh->dev;
        struct in_device *in_dev;
        struct neigh_parms *parms;
@@ -296,16 +295,19 @@ static int arp_constructor(struct neighbour *neigh)
                        neigh->ops = &arp_broken_ops;
                        neigh->output = neigh->ops->output;
                        return 0;
+#else
+                        break;
 #endif
-                ;}
+                }
 #endif
                if (neigh->type == RTN_MULTICAST) {
                        neigh->nud_state = NUD_NOARP;
                        arp_mc_map(addr, neigh->ha, dev, 1);
-                } else if (dev->flags&(IFF_NOARP|IFF_LOOPBACK)) {
+                } else if (dev->flags & (IFF_NOARP | IFF_LOOPBACK)) {
                        neigh->nud_state = NUD_NOARP;
                        memcpy(neigh->ha, dev->dev_addr, dev->addr_len);
-                } else if (neigh->type == RTN_BROADCAST || dev->flags&IFF_POINTOPOINT) {
+                } else if (neigh->type == RTN_BROADCAST ||
+                           (dev->flags & IFF_POINTOPOINT)) {
                        neigh->nud_state = NUD_NOARP;
                        memcpy(neigh->ha, dev->broadcast, dev->addr_len);
                }
@@ -315,7 +317,7 @@ static int arp_constructor(struct neighbour *neigh)
                else
                        neigh->ops = &arp_generic_ops;
-                if (neigh->nud_state&NUD_VALID)
+                if (neigh->nud_state & NUD_VALID)
                        neigh->output = neigh->ops->connected_output;
                else
                        neigh->output = neigh->ops->output;
@@ -334,7 +336,7 @@ static void arp_solicit(struct neighbour *neigh, struct sk_buff *skb)
        __be32 saddr = 0;
        u8  *dst_ha = NULL;
        struct net_device *dev = neigh->dev;
-        __be32 target = *(__be32*)neigh->primary_key;
+        __be32 target = *(__be32 *)neigh->primary_key;
        int probes = atomic_read(&neigh->probes);
        struct in_device *in_dev;
@@ -347,7 +349,8 @@ static void arp_solicit(struct neighbour *neigh, struct sk_buff *skb)
        switch (IN_DEV_ARP_ANNOUNCE(in_dev)) {
        default:
        case 0:         /* By default announce any local IP */
-                if (skb && inet_addr_type(dev_net(dev), ip_hdr(skb)->saddr) == RTN_LOCAL)
+                if (skb && inet_addr_type(dev_net(dev),
+                                          ip_hdr(skb)->saddr) == RTN_LOCAL)
                        saddr = ip_hdr(skb)->saddr;
                break;
        case 1:         /* Restrict announcements of saddr in same subnet */
@@ -369,16 +372,21 @@ static void arp_solicit(struct neighbour *neigh, struct sk_buff *skb)
        if (!saddr)
                saddr = inet_select_addr(dev, target, RT_SCOPE_LINK);
-        if ((probes -= neigh->parms->ucast_probes) < 0) {
+        probes -= neigh->parms->ucast_probes;
-                if (!(neigh->nud_state&NUD_VALID))
+        if (probes < 0) {
-                        printk(KERN_DEBUG "trying to ucast probe in NUD_INVALID\n");
+                if (!(neigh->nud_state & NUD_VALID))
+                        printk(KERN_DEBUG
+                               "trying to ucast probe in NUD_INVALID\n");
                dst_ha = neigh->ha;
                read_lock_bh(&neigh->lock);
-        } else if ((probes -= neigh->parms->app_probes) < 0) {
+        } else {
+                probes -= neigh->parms->app_probes;
+                if (probes < 0) {
 #ifdef CONFIG_ARPD
-                neigh_app_ns(neigh);
+                        neigh_app_ns(neigh);
 #endif
-                return;
+                        return;
+                }
        }
        arp_send(ARPOP_REQUEST, ETH_P_ARP, target, dev, saddr,
@@ -451,7 +459,8 @@ static int arp_filter(__be32 sip, __be32 tip, struct net_device *dev)
 *      is allowed to use this function, it is scheduled to be removed. --ANK
 */
-static int arp_set_predefined(int addr_hint, unsigned char * haddr, __be32 paddr, struct net_device * dev)
+static int arp_set_predefined(int addr_hint, unsigned char *haddr,
+                              __be32 paddr, struct net_device *dev)
 {
        switch (addr_hint) {
        case RTN_LOCAL:
@@ -483,7 +492,8 @@ int arp_find(unsigned char *haddr, struct sk_buff *skb)
        paddr = skb_rtable(skb)->rt_gateway;
-        if (arp_set_predefined(inet_addr_type(dev_net(dev), paddr), haddr, paddr, dev))
+        if (arp_set_predefined(inet_addr_type(dev_net(dev), paddr), haddr,
+                               paddr, dev))
                return 0;
        n = __neigh_lookup(&arp_tbl, &paddr, dev, 1);
@@ -515,13 +525,14 @@ int arp_bind_neighbour(struct dst_entry *dst)
                return -EINVAL;
        if (n == NULL) {
                __be32 nexthop = ((struct rtable *)dst)->rt_gateway;
-                if (dev->flags&(IFF_LOOPBACK|IFF_POINTOPOINT))
+                if (dev->flags & (IFF_LOOPBACK | IFF_POINTOPOINT))
                        nexthop = 0;
                n = __neigh_lookup_errno(
 #if defined(CONFIG_ATM_CLIP) || defined(CONFIG_ATM_CLIP_MODULE)
-                    dev->type == ARPHRD_ATM ? clip_tbl_hook :
+                                         dev->type == ARPHRD_ATM ?
+                                         clip_tbl_hook :
 #endif
-                    &arp_tbl, &nexthop, dev);
+                                         &arp_tbl, &nexthop, dev);
                if (IS_ERR(n))
                        return PTR_ERR(n);
                dst->neighbour = n;
@@ -543,8 +554,8 @@ static inline int arp_fwd_proxy(struct in_device *in_dev,
        if (!IN_DEV_PROXY_ARP(in_dev))
                return 0;
+        imi = IN_DEV_MEDIUM_ID(in_dev);
-        if ((imi = IN_DEV_MEDIUM_ID(in_dev)) == 0)
+        if (imi == 0)
                return 1;
        if (imi == -1)
                return 0;
@@ -555,7 +566,7 @@ static inline int arp_fwd_proxy(struct in_device *in_dev,
        if (out_dev)
                omi = IN_DEV_MEDIUM_ID(out_dev);
-        return (omi != imi && omi != -1);
+        return omi != imi && omi != -1;
 }
 /*
@@ -685,7 +696,7 @@ struct sk_buff *arp_create(int type, int ptype, __be32 dest_ip,
        arp->ar_pln = 4;
        arp->ar_op = htons(type);
-        arp_ptr=(unsigned char *)(arp+1);
+        arp_ptr = (unsigned char *)(arp + 1);
        memcpy(arp_ptr, src_hw, dev->addr_len);
        arp_ptr += dev->addr_len;
@@ -735,9 +746,8 @@ void arp_send(int type, int ptype, __be32 dest_ip,
        skb = arp_create(type, ptype, dest_ip, dev, src_ip,
                         dest_hw, src_hw, target_hw);
-        if (skb == NULL) {
+        if (skb == NULL)
                return;
-        }
        arp_xmit(skb);
 }
@@ -815,7 +825,7 @@ static int arp_process(struct sk_buff *skb)
 /*
 *      Extract fields
 */
-        arp_ptr= (unsigned char *)(arp+1);
+        arp_ptr = (unsigned char *)(arp + 1);
        sha     = arp_ptr;
        arp_ptr += dev->addr_len;
        memcpy(&sip, arp_ptr, 4);
@@ -869,16 +879,17 @@ static int arp_process(struct sk_buff *skb)
                addr_type = rt->rt_type;
                if (addr_type == RTN_LOCAL) {
-                        int dont_send = 0;
+                        int dont_send;
-                        if (!dont_send)
+                        dont_send = arp_ignore(in_dev, sip, tip);
-                                dont_send |= arp_ignore(in_dev,sip,tip);
                        if (!dont_send && IN_DEV_ARPFILTER(in_dev))
-                                dont_send |= arp_filter(sip,tip,dev);
+                                dont_send |= arp_filter(sip, tip, dev);
                        if (!dont_send) {
                                n = neigh_event_ns(&arp_tbl, sha, &sip, dev);
                                if (n) {
-                                        arp_send(ARPOP_REPLY,ETH_P_ARP,sip,dev,tip,sha,dev->dev_addr,sha);
+                                        arp_send(ARPOP_REPLY, ETH_P_ARP, sip,
+                                                 dev, tip, sha, dev->dev_addr,
+                                                 sha);
                                        neigh_release(n);
                                }
                        }
@@ -887,8 +898,7 @@ static int arp_process(struct sk_buff *skb)
                        if (addr_type == RTN_UNICAST  &&
                            (arp_fwd_proxy(in_dev, dev, rt) ||
                             arp_fwd_pvlan(in_dev, dev, rt, sip, tip) ||
-                             pneigh_lookup(&arp_tbl, net, &tip, dev, 0)))
+                             pneigh_lookup(&arp_tbl, net, &tip, dev, 0))) {
-                        {
                                n = neigh_event_ns(&arp_tbl, sha, &sip, dev);
                                if (n)
                                        neigh_release(n);
@@ -896,9 +906,12 @@ static int arp_process(struct sk_buff *skb)
                                if (NEIGH_CB(skb)->flags & LOCALLY_ENQUEUED ||
                                    skb->pkt_type == PACKET_HOST ||
                                    in_dev->arp_parms->proxy_delay == 0) {
-                                        arp_send(ARPOP_REPLY,ETH_P_ARP,sip,dev,tip,sha,dev->dev_addr,sha);
+                                        arp_send(ARPOP_REPLY, ETH_P_ARP, sip,
+                                                 dev, tip, sha, dev->dev_addr,
+                                                 sha);
                                } else {
-                                        pneigh_enqueue(&arp_tbl, in_dev->arp_parms, skb);
+                                        pneigh_enqueue(&arp_tbl,
+                                                       in_dev->arp_parms, skb);
                                        return 0;
                                }
                                goto out;
@@ -939,7 +952,8 @@ static int arp_process(struct sk_buff *skb)
                if (arp->ar_op != htons(ARPOP_REPLY) ||
                    skb->pkt_type != PACKET_HOST)
                        state = NUD_STALE;
-                neigh_update(n, sha, state, override ? NEIGH_UPDATE_F_OVERRIDE : 0);
+                neigh_update(n, sha, state,
+                             override ? NEIGH_UPDATE_F_OVERRIDE : 0);
                neigh_release(n);
        }
@@ -975,7 +989,8 @@ static int arp_rcv(struct sk_buff *skb, struct net_device *dev,
            arp->ar_pln != 4)
                goto freeskb;
-        if ((skb = skb_share_check(skb, GFP_ATOMIC)) == NULL)
+        skb = skb_share_check(skb, GFP_ATOMIC);
+        if (skb == NULL)
                goto out_of_mem;
        memset(NEIGH_CB(skb), 0, sizeof(struct neighbour_cb));
@@ -1019,7 +1034,7 @@ static int arp_req_set_public(struct net *net, struct arpreq *r,
                return -EINVAL;
        if (!dev && (r->arp_flags & ATF_COM)) {
                dev = dev_getbyhwaddr(net, r->arp_ha.sa_family,
-                                r->arp_ha.sa_data);
+                                      r->arp_ha.sa_data);
                if (!dev)
                        return -ENODEV;
        }
@@ -1033,7 +1048,7 @@ static int arp_req_set_public(struct net *net, struct arpreq *r,
 }
 static int arp_req_set(struct net *net, struct arpreq *r,
-                struct net_device * dev)
+                       struct net_device *dev)
 {
        __be32 ip;
        struct neighbour *neigh;
@@ -1046,10 +1061,11 @@ static int arp_req_set(struct net *net, struct arpreq *r,
        if (r->arp_flags & ATF_PERM)
                r->arp_flags |= ATF_COM;
        if (dev == NULL) {
-                struct flowi fl = { .nl_u = { .ip4_u = { .daddr = ip,
+                struct flowi fl = { .nl_u.ip4_u = { .daddr = ip,
-                                                         .tos = RTO_ONLINK } } };
+                                                    .tos = RTO_ONLINK } };
-                struct rtable * rt;
+                struct rtable *rt;
-                if ((err = ip_route_output_key(net, &rt, &fl)) != 0)
+                err = ip_route_output_key(net, &rt, &fl);
+                if (err != 0)
                        return err;
                dev = rt->dst.dev;
                ip_rt_put(rt);
@@ -1083,9 +1099,9 @@ static int arp_req_set(struct net *net, struct arpreq *r,
                unsigned state = NUD_STALE;
                if (r->arp_flags & ATF_PERM)
                        state = NUD_PERMANENT;
-                err = neigh_update(neigh, (r->arp_flags&ATF_COM) ?
+                err = neigh_update(neigh, (r->arp_flags & ATF_COM) ?
                                   r->arp_ha.sa_data : NULL, state,
-                                   NEIGH_UPDATE_F_OVERRIDE|
+                                   NEIGH_UPDATE_F_OVERRIDE |
                                   NEIGH_UPDATE_F_ADMIN);
                neigh_release(neigh);
        }
@@ -1094,12 +1110,12 @@ static int arp_req_set(struct net *net, struct arpreq *r,
 static unsigned arp_state_to_flags(struct neighbour *neigh)
 {
-        unsigned flags = 0;
        if (neigh->nud_state&NUD_PERMANENT)
-                flags = ATF_PERM|ATF_COM;
+                return ATF_PERM | ATF_COM;
        else if (neigh->nud_state&NUD_VALID)
-                flags = ATF_COM;
+                return ATF_COM;
-        return flags;
+        else
+                return 0;
 }
 /*
@@ -1142,7 +1158,7 @@ static int arp_req_delete_public(struct net *net, struct arpreq *r,
 }
 static int arp_req_delete(struct net *net, struct arpreq *r,
-                struct net_device * dev)
+                          struct net_device *dev)
 {
        int err;
        __be32 ip;
@@ -1153,10 +1169,11 @@ static int arp_req_delete(struct net *net, struct arpreq *r,
        ip = ((struct sockaddr_in *)&r->arp_pa)->sin_addr.s_addr;
        if (dev == NULL) {
-                struct flowi fl = { .nl_u = { .ip4_u = { .daddr = ip,
+                struct flowi fl = { .nl_u.ip4_u = { .daddr = ip,
-                                                         .tos = RTO_ONLINK } } };
+                                                    .tos = RTO_ONLINK } };
-                struct rtable * rt;
+                struct rtable *rt;
-                if ((err = ip_route_output_key(net, &rt, &fl)) != 0)
+                err = ip_route_output_key(net, &rt, &fl);
+                if (err != 0)
                        return err;
                dev = rt->dst.dev;
                ip_rt_put(rt);
@@ -1166,7 +1183,7 @@ static int arp_req_delete(struct net *net, struct arpreq *r,
        err = -ENXIO;
        neigh = neigh_lookup(&arp_tbl, &ip, dev);
        if (neigh) {
-                if (neigh->nud_state&~NUD_NOARP)
+                if (neigh->nud_state & ~NUD_NOARP)
                        err = neigh_update(neigh, NULL, NUD_FAILED,
                                           NEIGH_UPDATE_F_OVERRIDE|
                                           NEIGH_UPDATE_F_ADMIN);
@@ -1186,24 +1203,24 @@ int arp_ioctl(struct net *net, unsigned int cmd, void __user *arg)
        struct net_device *dev = NULL;
        switch (cmd) {
-                case SIOCDARP:
+        case SIOCDARP:
-                case SIOCSARP:
+        case SIOCSARP:
-                        if (!capable(CAP_NET_ADMIN))
+                if (!capable(CAP_NET_ADMIN))
-                                return -EPERM;
+                        return -EPERM;
-                case SIOCGARP:
+        case SIOCGARP:
-                        err = copy_from_user(&r, arg, sizeof(struct arpreq));
+                err = copy_from_user(&r, arg, sizeof(struct arpreq));
-                        if (err)
+                if (err)
-                                return -EFAULT;
+                        return -EFAULT;
-                        break;
+                break;
-                default:
+        default:
-                        return -EINVAL;
+                return -EINVAL;
        }
        if (r.arp_pa.sa_family != AF_INET)
                return -EPFNOSUPPORT;
        if (!(r.arp_flags & ATF_PUBL) &&
-            (r.arp_flags & (ATF_NETMASK|ATF_DONTPUB)))
+            (r.arp_flags & (ATF_NETMASK | ATF_DONTPUB)))
                return -EINVAL;
        if (!(r.arp_flags & ATF_NETMASK))
                ((struct sockaddr_in *)&r.arp_netmask)->sin_addr.s_addr =
@@ -1211,7 +1228,8 @@ int arp_ioctl(struct net *net, unsigned int cmd, void __user *arg)
        rtnl_lock();
        if (r.arp_dev[0]) {
                err = -ENODEV;
-                if ((dev = __dev_get_by_name(net, r.arp_dev)) == NULL)
+                dev = __dev_get_by_name(net, r.arp_dev);
+                if (dev == NULL)
                        goto out;
                /* Mmmm... It is wrong... ARPHRD_NETROM==0 */
@@ -1243,7 +1261,8 @@ out:
        return err;
 }
-static int arp_netdev_event(struct notifier_block *this, unsigned long event, void *ptr)
+static int arp_netdev_event(struct notifier_block *this, unsigned long event,
+                            void *ptr)
 {
        struct net_device *dev = ptr;
@@ -1311,12 +1330,13 @@ static char *ax2asc2(ax25_address *a, char *buf)
        for (n = 0, s = buf; n < 6; n++) {
                c = (a->ax25_call[n] >> 1) & 0x7F;
-                if (c != ' ') *s++ = c;
+                if (c != ' ')
+                        *s++ = c;
        }
        *s++ = '-';
+        n = (a->ax25_call[6] >> 1) & 0x0F;
-        if ((n = ((a->ax25_call[6] >> 1) & 0x0F)) > 9) {
+        if (n > 9) {
                *s++ = '1';
                n -= 10;
        }
@@ -1325,10 +1345,9 @@ static char *ax2asc2(ax25_address *a, char *buf)
        *s++ = '\0';
        if (*buf == '\0' || *buf == '-')
-           return "*";
+                return "*";
        return buf;
 }
 #endif /* CONFIG_AX25 */
@@ -1408,10 +1427,10 @@ static void *arp_seq_start(struct seq_file *seq, loff_t *pos)
 /* ------------------------------------------------------------------------ */
 static const struct seq_operations arp_seq_ops = {
-        .start  = arp_seq_start,
+        .start  = arp_seq_start,
-        .next   = neigh_seq_next,
+        .next   = neigh_seq_next,
-        .stop   = neigh_seq_stop,
+        .stop   = neigh_seq_stop,
-        .show   = arp_seq_show,
+        .show   = arp_seq_show,
 };
 static int arp_seq_open(struct inode *inode, struct file *file)
diff --git a/net/ipv4/datagram.c b/net/ipv4/datagram.c
index 721a8a37b45c..174be6caa5c8 100644
--- a/net/ipv4/datagram.c
+++ b/net/ipv4/datagram.c
@@ -73,6 +73,6 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)
        inet->inet_id = jiffies;
        sk_dst_set(sk, &rt->dst);
-        return(0);
+        return 0;
 }
 EXPORT_SYMBOL(ip4_datagram_connect);
diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c
index da14c49284f4..c2ff48fa18c7 100644
--- a/net/ipv4/devinet.c
+++ b/net/ipv4/devinet.c
@@ -209,7 +209,7 @@ static void inetdev_destroy(struct in_device *in_dev)
                inet_free_ifa(ifa);
        }
-        dev->ip_ptr = NULL;
+        rcu_assign_pointer(dev->ip_ptr, NULL);
        devinet_sysctl_unregister(in_dev);
        neigh_parms_release(&arp_tbl, in_dev->arp_parms);
@@ -1059,7 +1059,7 @@ static int inetdev_event(struct notifier_block *this, unsigned long event,
        switch (event) {
        case NETDEV_REGISTER:
                printk(KERN_DEBUG "inetdev_event: bug\n");
-                dev->ip_ptr = NULL;
+                rcu_assign_pointer(dev->ip_ptr, NULL);
                break;
        case NETDEV_UP:
                if (!inetdev_valid_mtu(dev->mtu))
diff --git a/net/ipv4/fib_frontend.c b/net/ipv4/fib_frontend.c
index 7d02a9f999fa..4a69a957872b 100644
--- a/net/ipv4/fib_frontend.c
+++ b/net/ipv4/fib_frontend.c
@@ -147,35 +147,40 @@ static void fib_flush(struct net *net)
                rt_cache_flush(net, -1);
 }
-/*
+/**
- *      Find the first device with a given source address.
+ * __ip_dev_find - find the first device with a given source address.
+ * @net: the net namespace
+ * @addr: the source address
+ * @devref: if true, take a reference on the found device
+ *
+ * If a caller uses devref=false, it should be protected by RCU
 */
+struct net_device *__ip_dev_find(struct net *net, __be32 addr, bool devref)
-struct net_device * ip_dev_find(struct net *net, __be32 addr)
 {
-        struct flowi fl = { .nl_u = { .ip4_u = { .daddr = addr } } };
+        struct flowi fl = {
-        struct fib_result res;
+                .nl_u = {
+                        .ip4_u = {
+                                .daddr = addr
+                        }
+                },
+                .flags = FLOWI_FLAG_MATCH_ANY_IIF
+        };
+        struct fib_result res = { 0 };
        struct net_device *dev = NULL;
-        struct fib_table *local_table;
-#ifdef CONFIG_IP_MULTIPLE_TABLES
+        if (fib_lookup(net, &fl, &res))
-        res.r = NULL;
-#endif
-        local_table = fib_get_table(net, RT_TABLE_LOCAL);
-        if (!local_table || fib_table_lookup(local_table, &fl, &res))
                return NULL;
        if (res.type != RTN_LOCAL)
                goto out;
        dev = FIB_RES_DEV(res);
-        if (dev)
+        if (dev && devref)
                dev_hold(dev);
 out:
        fib_res_put(&res);
        return dev;
 }
-EXPORT_SYMBOL(ip_dev_find);
+EXPORT_SYMBOL(__ip_dev_find);
 /*
 * Find address type as if only "dev" was present in the system. If
diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 4a8e370862bc..a96e5ec211a0 100644
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -186,9 +186,7 @@ static inline struct tnode *node_parent_rcu(struct node *node)
 {
        struct tnode *ret = node_parent(node);
-        return rcu_dereference_check(ret,
+        return rcu_dereference_rtnl(ret);
-                                     rcu_read_lock_held() ||
-                                     lockdep_rtnl_is_held());
 }
 /* Same as rcu_assign_pointer
@@ -211,9 +209,7 @@ static inline struct node *tnode_get_child_rcu(struct tnode *tn, unsigned int i)
 {
        struct node *ret = tnode_get_child(tn, i);
-        return rcu_dereference_check(ret,
+        return rcu_dereference_rtnl(ret);
-                                     rcu_read_lock_held() ||
-                                     lockdep_rtnl_is_held());
 }
 static inline int tnode_child_length(const struct tnode *tn)
@@ -459,8 +455,8 @@ static struct tnode *tnode_new(t_key key, int pos, int bits)
                tn->empty_children = 1<<bits;
        }
-        pr_debug("AT %p s=%u %lu\n", tn, (unsigned int) sizeof(struct tnode),
+        pr_debug("AT %p s=%zu %zu\n", tn, sizeof(struct tnode),
-                 (unsigned long) (sizeof(struct node) << bits));
+                 sizeof(struct node) << bits);
        return tn;
 }
@@ -609,11 +605,10 @@ static struct node *resize(struct trie *t, struct tnode *tn)
        /* Keep root node larger  */
-        if (!node_parent((struct node*) tn)) {
+        if (!node_parent((struct node *)tn)) {
                inflate_threshold_use = inflate_threshold_root;
                halve_threshold_use = halve_threshold_root;
-        }
+        } else {
-        else {
                inflate_threshold_use = inflate_threshold;
                halve_threshold_use = halve_threshold;
        }
@@ -639,7 +634,7 @@ static struct node *resize(struct trie *t, struct tnode *tn)
        check_tnode(tn);
        /* Return if at least one inflate is run */
-        if( max_work != MAX_WORK)
+        if (max_work != MAX_WORK)
                return (struct node *) tn;
        /*
@@ -966,9 +961,7 @@ fib_find_node(struct trie *t, u32 key)
        struct node *n;
        pos = 0;
-        n = rcu_dereference_check(t->trie,
+        n = rcu_dereference_rtnl(t->trie);
-                                  rcu_read_lock_held() ||
-                                  lockdep_rtnl_is_held());
        while (n != NULL &&  NODE_TYPE(n) == T_TNODE) {
                tn = (struct tnode *) n;
@@ -1748,16 +1741,14 @@ static struct leaf *leaf_walk_rcu(struct tnode *p, struct node *c)
                /* Node empty, walk back up to parent */
                c = (struct node *) p;
-        } while ( (p = node_parent_rcu(c)) != NULL);
+        } while ((p = node_parent_rcu(c)) != NULL);
        return NULL; /* Root of trie */
 }
 static struct leaf *trie_firstleaf(struct trie *t)
 {
-        struct tnode *n = (struct tnode *) rcu_dereference_check(t->trie,
+        struct tnode *n = (struct tnode *)rcu_dereference_rtnl(t->trie);
-                                                        rcu_read_lock_held() ||
-                                                        lockdep_rtnl_is_held());
        if (!n)
                return NULL;
@@ -2043,14 +2034,14 @@ struct fib_trie_iter {
        struct seq_net_private p;
        struct fib_table *tb;
        struct tnode *tnode;
-        unsigned index;
+        unsigned int index;
-        unsigned depth;
+        unsigned int depth;
 };
 static struct node *fib_trie_get_next(struct fib_trie_iter *iter)
 {
        struct tnode *tn = iter->tnode;
-        unsigned cindex = iter->index;
+        unsigned int cindex = iter->index;
        struct tnode *p;
        /* A single entry routing table */
@@ -2159,7 +2150,7 @@ static void trie_collect_stats(struct trie *t, struct trie_stat *s)
 */
 static void trie_show_stats(struct seq_file *seq, struct trie_stat *stat)
 {
-        unsigned i, max, pointers, bytes, avdepth;
+        unsigned int i, max, pointers, bytes, avdepth;
        if (stat->leaves)
                avdepth = stat->totdepth*100 / stat->leaves;
@@ -2356,7 +2347,8 @@ static void fib_trie_seq_stop(struct seq_file *seq, void *v)
 static void seq_indent(struct seq_file *seq, int n)
 {
-        while (n-- > 0) seq_puts(seq, "   ");
+        while (n-- > 0)
+                seq_puts(seq, "   ");
 }
 static inline const char *rtn_scope(char *buf, size_t len, enum rt_scope_t s)
@@ -2388,7 +2380,7 @@ static const char *const rtn_type_names[__RTN_MAX] = {
        [RTN_XRESOLVE] = "XRESOLVE",
 };
-static inline const char *rtn_type(char *buf, size_t len, unsigned t)
+static inline const char *rtn_type(char *buf, size_t len, unsigned int t)
 {
        if (t < __RTN_MAX && rtn_type_names[t])
                return rtn_type_names[t];
@@ -2544,13 +2536,12 @@ static void fib_route_seq_stop(struct seq_file *seq, void *v)
        rcu_read_unlock();
 }
-static unsigned fib_flag_trans(int type, __be32 mask, const struct fib_info *fi)
+static unsigned int fib_flag_trans(int type, __be32 mask, const struct fib_info *fi)
 {
-        static unsigned type2flags[RTN_MAX + 1] = {
+        unsigned int flags = 0;
-                [7] = RTF_REJECT, [8] = RTF_REJECT,
-        };
-        unsigned flags = type2flags[type];
+        if (type == RTN_UNREACHABLE || type == RTN_PROHIBIT)
+                flags = RTF_REJECT;
        if (fi && fi->fib_nh->nh_gw)
                flags |= RTF_GATEWAY;
        if (mask == htonl(0xFFFFFFFF))
@@ -2562,7 +2553,7 @@ static unsigned fib_flag_trans(int type, __be32 mask, const struct fib_info *fi)
 /*
 *      This outputs /proc/net/route.
 *      The format of the file is not supposed to be changed
- *      and needs to be same as fib_hash output to avoid breaking
+ *      and needs to be same as fib_hash output to avoid breaking
 *      legacy utilities
 */
 static int fib_route_seq_show(struct seq_file *seq, void *v)
@@ -2587,7 +2578,7 @@ static int fib_route_seq_show(struct seq_file *seq, void *v)
                list_for_each_entry_rcu(fa, &li->falh, fa_list) {
                        const struct fib_info *fi = fa->fa_info;
-                        unsigned flags = fib_flag_trans(fa->fa_type, mask, fi);
+                        unsigned int flags = fib_flag_trans(fa->fa_type, mask, fi);
                        int len;
                        if (fa->fa_type == RTN_BROADCAST
diff --git a/net/ipv4/gre.c b/net/ipv4/gre.c
new file mode 100644
index 000000000000..caea6885fdbd
--- /dev/null
+++ b/net/ipv4/gre.c
@@ -0,0 +1,151 @@
+/*
+ *      GRE over IPv4 demultiplexer driver
+ *
+ *      Authors: Dmitry Kozlov (xeb@mail.ru)
+ *
+ *      This program is free software; you can redistribute it and/or
+ *      modify it under the terms of the GNU General Public License
+ *      as published by the Free Software Foundation; either version
+ *      2 of the License, or (at your option) any later version.
+ *
+ */
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/kmod.h>
+#include <linux/skbuff.h>
+#include <linux/in.h>
+#include <linux/netdevice.h>
+#include <linux/version.h>
+#include <linux/spinlock.h>
+#include <net/protocol.h>
+#include <net/gre.h>
+static const struct gre_protocol *gre_proto[GREPROTO_MAX] __read_mostly;
+static DEFINE_SPINLOCK(gre_proto_lock);
+int gre_add_protocol(const struct gre_protocol *proto, u8 version)
+{
+        if (version >= GREPROTO_MAX)
+                goto err_out;
+        spin_lock(&gre_proto_lock);
+        if (gre_proto[version])
+                goto err_out_unlock;
+        rcu_assign_pointer(gre_proto[version], proto);
+        spin_unlock(&gre_proto_lock);
+        return 0;
+err_out_unlock:
+        spin_unlock(&gre_proto_lock);
+err_out:
+        return -1;
+}
+EXPORT_SYMBOL_GPL(gre_add_protocol);
+int gre_del_protocol(const struct gre_protocol *proto, u8 version)
+{
+        if (version >= GREPROTO_MAX)
+                goto err_out;
+        spin_lock(&gre_proto_lock);
+        if (gre_proto[version] != proto)
+                goto err_out_unlock;
+        rcu_assign_pointer(gre_proto[version], NULL);
+        spin_unlock(&gre_proto_lock);
+        synchronize_rcu();
+        return 0;
+err_out_unlock:
+        spin_unlock(&gre_proto_lock);
+err_out:
+        return -1;
+}
+EXPORT_SYMBOL_GPL(gre_del_protocol);
+static int gre_rcv(struct sk_buff *skb)
+{
+        const struct gre_protocol *proto;
+        u8 ver;
+        int ret;
+        if (!pskb_may_pull(skb, 12))
+                goto drop;
+        ver = skb->data[1]&0x7f;
+        if (ver >= GREPROTO_MAX)
+                goto drop;
+        rcu_read_lock();
+        proto = rcu_dereference(gre_proto[ver]);
+        if (!proto || !proto->handler)
+                goto drop_unlock;
+        ret = proto->handler(skb);
+        rcu_read_unlock();
+        return ret;
+drop_unlock:
+        rcu_read_unlock();
+drop:
+        kfree_skb(skb);
+        return NET_RX_DROP;
+}
+static void gre_err(struct sk_buff *skb, u32 info)
+{
+        const struct gre_protocol *proto;
+        u8 ver;
+        if (!pskb_may_pull(skb, 12))
+                goto drop;
+        ver = skb->data[1]&0x7f;
+        if (ver >= GREPROTO_MAX)
+                goto drop;
+        rcu_read_lock();
+        proto = rcu_dereference(gre_proto[ver]);
+        if (!proto || !proto->err_handler)
+                goto drop_unlock;
+        proto->err_handler(skb, info);
+        rcu_read_unlock();
+        return;
+drop_unlock:
+        rcu_read_unlock();
+drop:
+        kfree_skb(skb);
+}
+static const struct net_protocol net_gre_protocol = {
+        .handler     = gre_rcv,
+        .err_handler = gre_err,
+        .netns_ok    = 1,
+};
+static int __init gre_init(void)
+{
+        pr_info("GRE over IPv4 demultiplexor driver");
+        if (inet_add_protocol(&net_gre_protocol, IPPROTO_GRE) < 0) {
+                pr_err("gre: can't add protocol\n");
+                return -EAGAIN;
+        }
+        return 0;
+}
+static void __exit gre_exit(void)
+{
+        inet_del_protocol(&net_gre_protocol, IPPROTO_GRE);
+}
+module_init(gre_init);
+module_exit(gre_exit);
+MODULE_DESCRIPTION("GRE over IPv4 demultiplexer driver");
+MODULE_AUTHOR("D. Kozlov (xeb@mail.ru)");
+MODULE_LICENSE("GPL");
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index a0d847c7cba5..96bc7f9475a3 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -379,7 +379,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
        inet->tos = ip_hdr(skb)->tos;
        daddr = ipc.addr = rt->rt_src;
        ipc.opt = NULL;
-        ipc.shtx.flags = 0;
+        ipc.tx_flags = 0;
        if (icmp_param->replyopts.optlen) {
                ipc.opt = &icmp_param->replyopts;
                if (ipc.opt->srr)
@@ -538,7 +538,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
        inet_sk(sk)->tos = tos;
        ipc.addr = iph->saddr;
        ipc.opt = &icmp_param.replyopts;
-        ipc.shtx.flags = 0;
+        ipc.tx_flags = 0;
        {
                struct flowi fl = {
diff --git a/net/ipv4/inet_diag.c b/net/ipv4/inet_diag.c
index e5fa2ddce320..ba8042665849 100644
--- a/net/ipv4/inet_diag.c
+++ b/net/ipv4/inet_diag.c
@@ -425,7 +425,7 @@ static int inet_diag_bc_run(const void *bc, int len,
                        bc += op->no;
                }
        }
-        return (len == 0);
+        return len == 0;
 }
 static int valid_cc(const void *bc, int len, int cc)
diff --git a/net/ipv4/ip_fragment.c b/net/ipv4/ip_fragment.c
index b7c41654dde5..168440834ade 100644
--- a/net/ipv4/ip_fragment.c
+++ b/net/ipv4/ip_fragment.c
@@ -116,11 +116,11 @@ static int ip4_frag_match(struct inet_frag_queue *q, void *a)
        struct ip4_create_arg *arg = a;
        qp = container_of(q, struct ipq, q);
-        return (qp->id == arg->iph->id &&
+        return  qp->id == arg->iph->id &&
                        qp->saddr == arg->iph->saddr &&
                        qp->daddr == arg->iph->daddr &&
                        qp->protocol == arg->iph->protocol &&
-                        qp->user == arg->user);
+                        qp->user == arg->user;
 }
 /* Memory Tracking Functions. */
@@ -542,7 +542,7 @@ static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev,
        /* If the first fragment is fragmented itself, we split
         * it to two chunks: the first with data and paged part
         * and the second, holding only fragments. */
-        if (skb_has_frags(head)) {
+        if (skb_has_frag_list(head)) {
                struct sk_buff *clone;
                int i, plen = 0;
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 35c93e8b6a46..fbe2c473a06a 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -44,6 +44,7 @@
 #include <net/net_namespace.h>
 #include <net/netns/generic.h>
 #include <net/rtnetlink.h>
+#include <net/gre.h>
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
 #include <net/ipv6.h>
@@ -63,13 +64,13 @@
   We cannot track such dead loops during route installation,
   it is infeasible task. The most general solutions would be
   to keep skb->encapsulation counter (sort of local ttl),
-   and silently drop packet when it expires. It is the best
+   and silently drop packet when it expires. It is a good
   solution, but it supposes maintaing new variable in ALL
   skb, even if no tunneling is used.
-   Current solution: HARD_TX_LOCK lock breaks dead loops.
+   Current solution: xmit_recursion breaks dead loops. This is a percpu
+   counter, since when we enter the first ndo_xmit(), cpu migration is
+   forbidden. We force an exit if this counter reaches RECURSION_LIMIT
   2. Networking dead loops would not kill routers, but would really
   kill network. IP hop limit plays role of "t->recursion" in this case,
@@ -128,7 +129,7 @@ static int ipgre_tunnel_bind_dev(struct net_device *dev);
 static int ipgre_net_id __read_mostly;
 struct ipgre_net {
-        struct ip_tunnel *tunnels[4][HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels[4][HASH_SIZE];
        struct net_device *fb_tunnel_dev;
 };
@@ -158,13 +159,40 @@ struct ipgre_net {
 #define tunnels_l       tunnels[1]
 #define tunnels_wc      tunnels[0]
 /*
- * Locking : hash tables are protected by RCU and a spinlock
+ * Locking : hash tables are protected by RCU and RTNL
 */
-static DEFINE_SPINLOCK(ipgre_lock);
 #define for_each_ip_tunnel_rcu(start) \
        for (t = rcu_dereference(start); t; t = rcu_dereference(t->next))
+/* often modified stats are per cpu, other are shared (netdev->stats) */
+struct pcpu_tstats {
+        unsigned long   rx_packets;
+        unsigned long   rx_bytes;
+        unsigned long   tx_packets;
+        unsigned long   tx_bytes;
+};
+static struct net_device_stats *ipgre_get_stats(struct net_device *dev)
+{
+        struct pcpu_tstats sum = { 0 };
+        int i;
+        for_each_possible_cpu(i) {
+                const struct pcpu_tstats *tstats = per_cpu_ptr(dev->tstats, i);
+                sum.rx_packets += tstats->rx_packets;
+                sum.rx_bytes   += tstats->rx_bytes;
+                sum.tx_packets += tstats->tx_packets;
+                sum.tx_bytes   += tstats->tx_bytes;
+        }
+        dev->stats.rx_packets = sum.rx_packets;
+        dev->stats.rx_bytes   = sum.rx_bytes;
+        dev->stats.tx_packets = sum.tx_packets;
+        dev->stats.tx_bytes   = sum.tx_bytes;
+        return &dev->stats;
+}
 /* Given src, dst and key, find appropriate for input tunnel. */
 static struct ip_tunnel * ipgre_tunnel_lookup(struct net_device *dev,
@@ -173,8 +201,8 @@ static struct ip_tunnel * ipgre_tunnel_lookup(struct net_device *dev,
 {
        struct net *net = dev_net(dev);
        int link = dev->ifindex;
-        unsigned h0 = HASH(remote);
+        unsigned int h0 = HASH(remote);
-        unsigned h1 = HASH(key);
+        unsigned int h1 = HASH(key);
        struct ip_tunnel *t, *cand = NULL;
        struct ipgre_net *ign = net_generic(net, ipgre_net_id);
        int dev_type = (gre_proto == htons(ETH_P_TEB)) ?
@@ -289,13 +317,13 @@ static struct ip_tunnel * ipgre_tunnel_lookup(struct net_device *dev,
        return NULL;
 }
-static struct ip_tunnel **__ipgre_bucket(struct ipgre_net *ign,
+static struct ip_tunnel __rcu **__ipgre_bucket(struct ipgre_net *ign,
                struct ip_tunnel_parm *parms)
 {
        __be32 remote = parms->iph.daddr;
        __be32 local = parms->iph.saddr;
        __be32 key = parms->i_key;
-        unsigned h = HASH(key);
+        unsigned int h = HASH(key);
        int prio = 0;
        if (local)
@@ -308,7 +336,7 @@ static struct ip_tunnel **__ipgre_bucket(struct ipgre_net *ign,
        return &ign->tunnels[prio][h];
 }
-static inline struct ip_tunnel **ipgre_bucket(struct ipgre_net *ign,
+static inline struct ip_tunnel __rcu **ipgre_bucket(struct ipgre_net *ign,
                struct ip_tunnel *t)
 {
        return __ipgre_bucket(ign, &t->parms);
@@ -316,23 +344,22 @@ static inline struct ip_tunnel **ipgre_bucket(struct ipgre_net *ign,
 static void ipgre_tunnel_link(struct ipgre_net *ign, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp = ipgre_bucket(ign, t);
+        struct ip_tunnel __rcu **tp = ipgre_bucket(ign, t);
-        spin_lock_bh(&ipgre_lock);
+        rcu_assign_pointer(t->next, rtnl_dereference(*tp));
-        t->next = *tp;
        rcu_assign_pointer(*tp, t);
-        spin_unlock_bh(&ipgre_lock);
 }
 static void ipgre_tunnel_unlink(struct ipgre_net *ign, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp;
+        struct ip_tunnel __rcu **tp;
+        struct ip_tunnel *iter;
-        for (tp = ipgre_bucket(ign, t); *tp; tp = &(*tp)->next) {
-                if (t == *tp) {
+        for (tp = ipgre_bucket(ign, t);
-                        spin_lock_bh(&ipgre_lock);
+             (iter = rtnl_dereference(*tp)) != NULL;
-                        *tp = t->next;
+             tp = &iter->next) {
-                        spin_unlock_bh(&ipgre_lock);
+                if (t == iter) {
+                        rcu_assign_pointer(*tp, t->next);
                        break;
                }
        }
@@ -346,10 +373,13 @@ static struct ip_tunnel *ipgre_tunnel_find(struct net *net,
        __be32 local = parms->iph.saddr;
        __be32 key = parms->i_key;
        int link = parms->link;
-        struct ip_tunnel *t, **tp;
+        struct ip_tunnel *t;
+        struct ip_tunnel __rcu **tp;
        struct ipgre_net *ign = net_generic(net, ipgre_net_id);
-        for (tp = __ipgre_bucket(ign, parms); (t = *tp) != NULL; tp = &t->next)
+        for (tp = __ipgre_bucket(ign, parms);
+             (t = rtnl_dereference(*tp)) != NULL;
+             tp = &t->next)
                if (local == t->parms.iph.saddr &&
                    remote == t->parms.iph.daddr &&
                    key == t->parms.i_key &&
@@ -360,7 +390,7 @@ static struct ip_tunnel *ipgre_tunnel_find(struct net *net,
        return t;
 }
-static struct ip_tunnel * ipgre_tunnel_locate(struct net *net,
+static struct ip_tunnel *ipgre_tunnel_locate(struct net *net,
                struct ip_tunnel_parm *parms, int create)
 {
        struct ip_tunnel *t, *nt;
@@ -582,7 +612,7 @@ static int ipgre_rcv(struct sk_buff *skb)
        if ((tunnel = ipgre_tunnel_lookup(skb->dev,
                                          iph->saddr, iph->daddr, key,
                                          gre_proto))) {
-                struct net_device_stats *stats = &tunnel->dev->stats;
+                struct pcpu_tstats *tstats;
                secpath_reset(skb);
@@ -606,22 +636,22 @@ static int ipgre_rcv(struct sk_buff *skb)
                        /* Looped back packet, drop it! */
                        if (skb_rtable(skb)->fl.iif == 0)
                                goto drop;
-                        stats->multicast++;
+                        tunnel->dev->stats.multicast++;
                        skb->pkt_type = PACKET_BROADCAST;
                }
 #endif
                if (((flags&GRE_CSUM) && csum) ||
                    (!(flags&GRE_CSUM) && tunnel->parms.i_flags&GRE_CSUM)) {
-                        stats->rx_crc_errors++;
+                        tunnel->dev->stats.rx_crc_errors++;
-                        stats->rx_errors++;
+                        tunnel->dev->stats.rx_errors++;
                        goto drop;
                }
                if (tunnel->parms.i_flags&GRE_SEQ) {
                        if (!(flags&GRE_SEQ) ||
                            (tunnel->i_seqno && (s32)(seqno - tunnel->i_seqno) < 0)) {
-                                stats->rx_fifo_errors++;
+                                tunnel->dev->stats.rx_fifo_errors++;
-                                stats->rx_errors++;
+                                tunnel->dev->stats.rx_errors++;
                                goto drop;
                        }
                        tunnel->i_seqno = seqno + 1;
@@ -630,8 +660,8 @@ static int ipgre_rcv(struct sk_buff *skb)
                /* Warning: All skb pointers will be invalidated! */
                if (tunnel->dev->type == ARPHRD_ETHER) {
                        if (!pskb_may_pull(skb, ETH_HLEN)) {
-                                stats->rx_length_errors++;
+                                tunnel->dev->stats.rx_length_errors++;
-                                stats->rx_errors++;
+                                tunnel->dev->stats.rx_errors++;
                                goto drop;
                        }
@@ -640,14 +670,20 @@ static int ipgre_rcv(struct sk_buff *skb)
                        skb_postpull_rcsum(skb, eth_hdr(skb), ETH_HLEN);
                }
-                skb_tunnel_rx(skb, tunnel->dev);
+                tstats = this_cpu_ptr(tunnel->dev->tstats);
+                tstats->rx_packets++;
+                tstats->rx_bytes += skb->len;
+                __skb_tunnel_rx(skb, tunnel->dev);
                skb_reset_network_header(skb);
                ipgre_ecn_decapsulate(iph, skb);
-                netif_rx(skb);
+                if (netif_rx(skb) == NET_RX_DROP)
+                        tunnel->dev->stats.rx_dropped++;
                rcu_read_unlock();
-                return(0);
+                return 0;
        }
        icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PORT_UNREACH, 0);
@@ -655,20 +691,19 @@ drop:
        rcu_read_unlock();
 drop_nolock:
        kfree_skb(skb);
-        return(0);
+        return 0;
 }
 static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
-        struct net_device_stats *stats = &dev->stats;
+        struct pcpu_tstats *tstats;
-        struct netdev_queue *txq = netdev_get_tx_queue(dev, 0);
        struct iphdr  *old_iph = ip_hdr(skb);
        struct iphdr  *tiph;
        u8     tos;
        __be16 df;
        struct rtable *rt;                      /* Route to the other host */
-        struct net_device *tdev;                        /* Device to other host */
+        struct net_device *tdev;                /* Device to other host */
        struct iphdr  *iph;                     /* Our new IP header */
        unsigned int max_headroom;              /* The extra header space needed */
        int    gre_hlen;
@@ -690,7 +725,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
                /* NBMA tunnel */
                if (skb_dst(skb) == NULL) {
-                        stats->tx_fifo_errors++;
+                        dev->stats.tx_fifo_errors++;
                        goto tx_error;
                }
@@ -736,14 +771,20 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
        }
        {
-                struct flowi fl = { .oif = tunnel->parms.link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = tunnel->parms.link,
-                                              { .daddr = dst,
+                        .nl_u = {
-                                                .saddr = tiph->saddr,
+                                .ip4_u = {
-                                                .tos = RT_TOS(tos) } },
+                                        .daddr = dst,
-                                    .proto = IPPROTO_GRE };
+                                        .saddr = tiph->saddr,
+                                        .tos = RT_TOS(tos)
+                                }
+                        },
+                        .proto = IPPROTO_GRE
+                }
+;
                if (ip_route_output_key(dev_net(dev), &rt, &fl)) {
-                        stats->tx_carrier_errors++;
+                        dev->stats.tx_carrier_errors++;
                        goto tx_error;
                }
        }
@@ -751,7 +792,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
        if (tdev == dev) {
                ip_rt_put(rt);
-                stats->collisions++;
+                dev->stats.collisions++;
                goto tx_error;
        }
@@ -814,7 +855,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
                        dev->needed_headroom = max_headroom;
                if (!new_skb) {
                        ip_rt_put(rt);
-                        txq->tx_dropped++;
+                        dev->stats.tx_dropped++;
                        dev_kfree_skb(skb);
                        return NETDEV_TX_OK;
                }
@@ -881,15 +922,15 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
        }
        nf_reset(skb);
+        tstats = this_cpu_ptr(dev->tstats);
-        IPTUNNEL_XMIT();
+        __IPTUNNEL_XMIT(tstats, &dev->stats);
        return NETDEV_TX_OK;
 tx_error_icmp:
        dst_link_failure(skb);
 tx_error:
-        stats->tx_errors++;
+        dev->stats.tx_errors++;
        dev_kfree_skb(skb);
        return NETDEV_TX_OK;
 }
@@ -909,13 +950,19 @@ static int ipgre_tunnel_bind_dev(struct net_device *dev)
        /* Guess output device to choose reasonable mtu and needed_headroom */
        if (iph->daddr) {
-                struct flowi fl = { .oif = tunnel->parms.link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = tunnel->parms.link,
-                                              { .daddr = iph->daddr,
+                        .nl_u = {
-                                                .saddr = iph->saddr,
+                                .ip4_u = {
-                                                .tos = RT_TOS(iph->tos) } },
+                                        .daddr = iph->daddr,
-                                    .proto = IPPROTO_GRE };
+                                        .saddr = iph->saddr,
+                                        .tos = RT_TOS(iph->tos)
+                                }
+                        },
+                        .proto = IPPROTO_GRE
+                };
                struct rtable *rt;
                if (!ip_route_output_key(dev_net(dev), &rt, &fl)) {
                        tdev = rt->dst.dev;
                        ip_rt_put(rt);
@@ -1012,7 +1059,7 @@ ipgre_tunnel_ioctl (struct net_device *dev, struct ifreq *ifr, int cmd)
                                        break;
                                }
                        } else {
-                                unsigned nflags = 0;
+                                unsigned int nflags = 0;
                                t = netdev_priv(dev);
@@ -1125,7 +1172,7 @@ static int ipgre_tunnel_change_mtu(struct net_device *dev, int new_mtu)
 static int ipgre_header(struct sk_buff *skb, struct net_device *dev,
                        unsigned short type,
-                        const void *daddr, const void *saddr, unsigned len)
+                        const void *daddr, const void *saddr, unsigned int len)
 {
        struct ip_tunnel *t = netdev_priv(dev);
        struct iphdr *iph = (struct iphdr *)skb_push(skb, t->hlen);
@@ -1167,13 +1214,19 @@ static int ipgre_open(struct net_device *dev)
        struct ip_tunnel *t = netdev_priv(dev);
        if (ipv4_is_multicast(t->parms.iph.daddr)) {
-                struct flowi fl = { .oif = t->parms.link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = t->parms.link,
-                                              { .daddr = t->parms.iph.daddr,
+                        .nl_u = {
-                                                .saddr = t->parms.iph.saddr,
+                                .ip4_u = {
-                                                .tos = RT_TOS(t->parms.iph.tos) } },
+                                        .daddr = t->parms.iph.daddr,
-                                    .proto = IPPROTO_GRE };
+                                        .saddr = t->parms.iph.saddr,
+                                        .tos = RT_TOS(t->parms.iph.tos)
+                                }
+                        },
+                        .proto = IPPROTO_GRE
+                };
                struct rtable *rt;
                if (ip_route_output_key(dev_net(dev), &rt, &fl))
                        return -EADDRNOTAVAIL;
                dev = rt->dst.dev;
@@ -1213,12 +1266,19 @@ static const struct net_device_ops ipgre_netdev_ops = {
        .ndo_start_xmit         = ipgre_tunnel_xmit,
        .ndo_do_ioctl           = ipgre_tunnel_ioctl,
        .ndo_change_mtu         = ipgre_tunnel_change_mtu,
+        .ndo_get_stats          = ipgre_get_stats,
 };
+static void ipgre_dev_free(struct net_device *dev)
+{
+        free_percpu(dev->tstats);
+        free_netdev(dev);
+}
 static void ipgre_tunnel_setup(struct net_device *dev)
 {
        dev->netdev_ops         = &ipgre_netdev_ops;
-        dev->destructor         = free_netdev;
+        dev->destructor         = ipgre_dev_free;
        dev->type               = ARPHRD_IPGRE;
        dev->needed_headroom    = LL_MAX_HEADER + sizeof(struct iphdr) + 4;
@@ -1256,6 +1316,10 @@ static int ipgre_tunnel_init(struct net_device *dev)
        } else
                dev->header_ops = &ipgre_header_ops;
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
        return 0;
 }
@@ -1274,14 +1338,13 @@ static void ipgre_fb_tunnel_init(struct net_device *dev)
        tunnel->hlen            = sizeof(struct iphdr) + 4;
        dev_hold(dev);
-        ign->tunnels_wc[0]      = tunnel;
+        rcu_assign_pointer(ign->tunnels_wc[0], tunnel);
 }
-static const struct net_protocol ipgre_protocol = {
+static const struct gre_protocol ipgre_protocol = {
-        .handler        =       ipgre_rcv,
+        .handler     = ipgre_rcv,
-        .err_handler    =       ipgre_err,
+        .err_handler = ipgre_err,
-        .netns_ok       =       1,
 };
 static void ipgre_destroy_tunnels(struct ipgre_net *ign, struct list_head *head)
@@ -1291,11 +1354,13 @@ static void ipgre_destroy_tunnels(struct ipgre_net *ign, struct list_head *head)
        for (prio = 0; prio < 4; prio++) {
                int h;
                for (h = 0; h < HASH_SIZE; h++) {
-                        struct ip_tunnel *t = ign->tunnels[prio][h];
+                        struct ip_tunnel *t;
+                        t = rtnl_dereference(ign->tunnels[prio][h]);
                        while (t != NULL) {
                                unregister_netdevice_queue(t->dev, head);
-                                t = t->next;
+                                t = rtnl_dereference(t->next);
                        }
                }
        }
@@ -1441,6 +1506,10 @@ static int ipgre_tap_init(struct net_device *dev)
        ipgre_tunnel_bind_dev(dev);
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
        return 0;
 }
@@ -1451,6 +1520,7 @@ static const struct net_device_ops ipgre_tap_netdev_ops = {
        .ndo_set_mac_address    = eth_mac_addr,
        .ndo_validate_addr      = eth_validate_addr,
        .ndo_change_mtu         = ipgre_tunnel_change_mtu,
+        .ndo_get_stats          = ipgre_get_stats,
 };
 static void ipgre_tap_setup(struct net_device *dev)
@@ -1459,7 +1529,7 @@ static void ipgre_tap_setup(struct net_device *dev)
        ether_setup(dev);
        dev->netdev_ops         = &ipgre_tap_netdev_ops;
-        dev->destructor         = free_netdev;
+        dev->destructor         = ipgre_dev_free;
        dev->iflink             = 0;
        dev->features           |= NETIF_F_NETNS_LOCAL;
@@ -1487,6 +1557,10 @@ static int ipgre_newlink(struct net *src_net, struct net_device *dev, struct nla
        if (!tb[IFLA_MTU])
                dev->mtu = mtu;
+        /* Can use a lockless transmit, unless we generate output sequences */
+        if (!(nt->parms.o_flags & GRE_SEQ))
+                dev->features |= NETIF_F_LLTX;
        err = register_netdevice(dev);
        if (err)
                goto out;
@@ -1522,7 +1596,7 @@ static int ipgre_changelink(struct net_device *dev, struct nlattr *tb[],
                t = nt;
                if (dev->type != ARPHRD_ETHER) {
-                        unsigned nflags = 0;
+                        unsigned int nflags = 0;
                        if (ipv4_is_multicast(p.iph.daddr))
                                nflags = IFF_BROADCAST;
@@ -1663,7 +1737,7 @@ static int __init ipgre_init(void)
        if (err < 0)
                return err;
-        err = inet_add_protocol(&ipgre_protocol, IPPROTO_GRE);
+        err = gre_add_protocol(&ipgre_protocol, GREPROTO_CISCO);
        if (err < 0) {
                printk(KERN_INFO "ipgre init: can't add protocol\n");
                goto add_proto_failed;
@@ -1683,7 +1757,7 @@ out:
 tap_ops_failed:
        rtnl_link_unregister(&ipgre_link_ops);
 rtnl_link_failed:
-        inet_del_protocol(&ipgre_protocol, IPPROTO_GRE);
+        gre_del_protocol(&ipgre_protocol, GREPROTO_CISCO);
 add_proto_failed:
        unregister_pernet_device(&ipgre_net_ops);
        goto out;
@@ -1693,7 +1767,7 @@ static void __exit ipgre_fini(void)
 {
        rtnl_link_unregister(&ipgre_tap_ops);
        rtnl_link_unregister(&ipgre_link_ops);
-        if (inet_del_protocol(&ipgre_protocol, IPPROTO_GRE) < 0)
+        if (gre_del_protocol(&ipgre_protocol, GREPROTO_CISCO) < 0)
                printk(KERN_INFO "ipgre close: can't remove protocol\n");
        unregister_pernet_device(&ipgre_net_ops);
 }
diff --git a/net/ipv4/ip_options.c b/net/ipv4/ip_options.c
index ba9836c488ed..1906fa35860c 100644
--- a/net/ipv4/ip_options.c
+++ b/net/ipv4/ip_options.c
@@ -466,7 +466,7 @@ error:
        }
        return -EINVAL;
 }
+EXPORT_SYMBOL(ip_options_compile);
 /*
 *      Undo all the changes done by ip_options_compile().
@@ -646,3 +646,4 @@ int ip_options_rcv_srr(struct sk_buff *skb)
        }
        return 0;
 }
+EXPORT_SYMBOL(ip_options_rcv_srr);
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c
index 7649d7750075..439d2a34ee44 100644
--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -487,7 +487,7 @@ int ip_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *))
         * LATER: this step can be merged to real generation of fragments,
         * we can switch to copy when see the first bad fragment.
         */
-        if (skb_has_frags(skb)) {
+        if (skb_has_frag_list(skb)) {
                struct sk_buff *frag, *frag2;
                int first_len = skb_pagelen(skb);
@@ -844,10 +844,9 @@ int ip_append_data(struct sock *sk,
                inet->cork.length = 0;
                sk->sk_sndmsg_page = NULL;
                sk->sk_sndmsg_off = 0;
-                if ((exthdrlen = rt->dst.header_len) != 0) {
+                exthdrlen = rt->dst.header_len;
-                        length += exthdrlen;
+                length += exthdrlen;
-                        transhdrlen += exthdrlen;
+                transhdrlen += exthdrlen;
-                }
        } else {
                rt = (struct rtable *)inet->cork.dst;
                if (inet->cork.flags & IPCORK_OPT)
@@ -934,16 +933,19 @@ alloc_new_skb:
                            !(rt->dst.dev->features&NETIF_F_SG))
                                alloclen = mtu;
                        else
-                                alloclen = datalen + fragheaderlen;
+                                alloclen = fraglen;
                        /* The last fragment gets additional space at tail.
                         * Note, with MSG_MORE we overallocate on fragments,
                         * because we have no idea what fragment will be
                         * the last.
                         */
-                        if (datalen == length + fraggap)
+                        if (datalen == length + fraggap) {
                                alloclen += rt->dst.trailer_len;
+                                /* make sure mtu is not reached */
+                                if (datalen > mtu - fragheaderlen - rt->dst.trailer_len)
+                                        datalen -= ALIGN(rt->dst.trailer_len, 8);
+                        }
                        if (transhdrlen) {
                                skb = sock_alloc_send_skb(sk,
                                                alloclen + hh_len + 15,
@@ -960,7 +962,7 @@ alloc_new_skb:
                                else
                                        /* only the initial fragment is
                                           time stamped */
-                                        ipc->shtx.flags = 0;
+                                        ipc->tx_flags = 0;
                        }
                        if (skb == NULL)
                                goto error;
@@ -971,7 +973,7 @@ alloc_new_skb:
                        skb->ip_summed = csummode;
                        skb->csum = 0;
                        skb_reserve(skb, hh_len);
-                        *skb_tx(skb) = ipc->shtx;
+                        skb_shinfo(skb)->tx_flags = ipc->tx_flags;
                        /*
                         *      Find where to start putting bytes.
@@ -1391,7 +1393,7 @@ void ip_send_reply(struct sock *sk, struct sk_buff *skb, struct ip_reply_arg *ar
        daddr = ipc.addr = rt->rt_src;
        ipc.opt = NULL;
-        ipc.shtx.flags = 0;
+        ipc.tx_flags = 0;
        if (replyopts.opt.optlen) {
                ipc.opt = &replyopts.opt;
diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c
index ec036731a70b..6ad46c28ede2 100644
--- a/net/ipv4/ipip.c
+++ b/net/ipv4/ipip.c
@@ -122,31 +122,59 @@
 static int ipip_net_id __read_mostly;
 struct ipip_net {
-        struct ip_tunnel *tunnels_r_l[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_r_l[HASH_SIZE];
-        struct ip_tunnel *tunnels_r[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_r[HASH_SIZE];
-        struct ip_tunnel *tunnels_l[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_l[HASH_SIZE];
-        struct ip_tunnel *tunnels_wc[1];
+        struct ip_tunnel __rcu *tunnels_wc[1];
-        struct ip_tunnel **tunnels[4];
+        struct ip_tunnel __rcu **tunnels[4];
        struct net_device *fb_tunnel_dev;
 };
-static void ipip_tunnel_init(struct net_device *dev);
+static int ipip_tunnel_init(struct net_device *dev);
 static void ipip_tunnel_setup(struct net_device *dev);
+static void ipip_dev_free(struct net_device *dev);
 /*
- * Locking : hash tables are protected by RCU and a spinlock
+ * Locking : hash tables are protected by RCU and RTNL
 */
-static DEFINE_SPINLOCK(ipip_lock);
 #define for_each_ip_tunnel_rcu(start) \
        for (t = rcu_dereference(start); t; t = rcu_dereference(t->next))
+/* often modified stats are per cpu, other are shared (netdev->stats) */
+struct pcpu_tstats {
+        unsigned long   rx_packets;
+        unsigned long   rx_bytes;
+        unsigned long   tx_packets;
+        unsigned long   tx_bytes;
+};
+static struct net_device_stats *ipip_get_stats(struct net_device *dev)
+{
+        struct pcpu_tstats sum = { 0 };
+        int i;
+        for_each_possible_cpu(i) {
+                const struct pcpu_tstats *tstats = per_cpu_ptr(dev->tstats, i);
+                sum.rx_packets += tstats->rx_packets;
+                sum.rx_bytes   += tstats->rx_bytes;
+                sum.tx_packets += tstats->tx_packets;
+                sum.tx_bytes   += tstats->tx_bytes;
+        }
+        dev->stats.rx_packets = sum.rx_packets;
+        dev->stats.rx_bytes   = sum.rx_bytes;
+        dev->stats.tx_packets = sum.tx_packets;
+        dev->stats.tx_bytes   = sum.tx_bytes;
+        return &dev->stats;
+}
 static struct ip_tunnel * ipip_tunnel_lookup(struct net *net,
                __be32 remote, __be32 local)
 {
-        unsigned h0 = HASH(remote);
+        unsigned int h0 = HASH(remote);
-        unsigned h1 = HASH(local);
+        unsigned int h1 = HASH(local);
        struct ip_tunnel *t;
        struct ipip_net *ipn = net_generic(net, ipip_net_id);
@@ -169,12 +197,12 @@ static struct ip_tunnel * ipip_tunnel_lookup(struct net *net,
        return NULL;
 }
-static struct ip_tunnel **__ipip_bucket(struct ipip_net *ipn,
+static struct ip_tunnel __rcu **__ipip_bucket(struct ipip_net *ipn,
                struct ip_tunnel_parm *parms)
 {
        __be32 remote = parms->iph.daddr;
        __be32 local = parms->iph.saddr;
-        unsigned h = 0;
+        unsigned int h = 0;
        int prio = 0;
        if (remote) {
@@ -188,7 +216,7 @@ static struct ip_tunnel **__ipip_bucket(struct ipip_net *ipn,
        return &ipn->tunnels[prio][h];
 }
-static inline struct ip_tunnel **ipip_bucket(struct ipip_net *ipn,
+static inline struct ip_tunnel __rcu **ipip_bucket(struct ipip_net *ipn,
                struct ip_tunnel *t)
 {
        return __ipip_bucket(ipn, &t->parms);
@@ -196,13 +224,14 @@ static inline struct ip_tunnel **ipip_bucket(struct ipip_net *ipn,
 static void ipip_tunnel_unlink(struct ipip_net *ipn, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp;
+        struct ip_tunnel __rcu **tp;
+        struct ip_tunnel *iter;
-        for (tp = ipip_bucket(ipn, t); *tp; tp = &(*tp)->next) {
-                if (t == *tp) {
+        for (tp = ipip_bucket(ipn, t);
-                        spin_lock_bh(&ipip_lock);
+             (iter = rtnl_dereference(*tp)) != NULL;
-                        *tp = t->next;
+             tp = &iter->next) {
-                        spin_unlock_bh(&ipip_lock);
+                if (t == iter) {
+                        rcu_assign_pointer(*tp, t->next);
                        break;
                }
        }
@@ -210,12 +239,10 @@ static void ipip_tunnel_unlink(struct ipip_net *ipn, struct ip_tunnel *t)
 static void ipip_tunnel_link(struct ipip_net *ipn, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp = ipip_bucket(ipn, t);
+        struct ip_tunnel __rcu **tp = ipip_bucket(ipn, t);
-        spin_lock_bh(&ipip_lock);
+        rcu_assign_pointer(t->next, rtnl_dereference(*tp));
-        t->next = *tp;
        rcu_assign_pointer(*tp, t);
-        spin_unlock_bh(&ipip_lock);
 }
 static struct ip_tunnel * ipip_tunnel_locate(struct net *net,
@@ -223,12 +250,15 @@ static struct ip_tunnel * ipip_tunnel_locate(struct net *net,
 {
        __be32 remote = parms->iph.daddr;
        __be32 local = parms->iph.saddr;
-        struct ip_tunnel *t, **tp, *nt;
+        struct ip_tunnel *t, *nt;
+        struct ip_tunnel __rcu **tp;
        struct net_device *dev;
        char name[IFNAMSIZ];
        struct ipip_net *ipn = net_generic(net, ipip_net_id);
-        for (tp = __ipip_bucket(ipn, parms); (t = *tp) != NULL; tp = &t->next) {
+        for (tp = __ipip_bucket(ipn, parms);
+                 (t = rtnl_dereference(*tp)) != NULL;
+                 tp = &t->next) {
                if (local == t->parms.iph.saddr && remote == t->parms.iph.daddr)
                        return t;
        }
@@ -238,7 +268,7 @@ static struct ip_tunnel * ipip_tunnel_locate(struct net *net,
        if (parms->name[0])
                strlcpy(name, parms->name, IFNAMSIZ);
        else
-                sprintf(name, "tunl%%d");
+                strcpy(name, "tunl%d");
        dev = alloc_netdev(sizeof(*t), name, ipip_tunnel_setup);
        if (dev == NULL)
@@ -254,7 +284,8 @@ static struct ip_tunnel * ipip_tunnel_locate(struct net *net,
        nt = netdev_priv(dev);
        nt->parms = *parms;
-        ipip_tunnel_init(dev);
+        if (ipip_tunnel_init(dev) < 0)
+                goto failed_free;
        if (register_netdevice(dev) < 0)
                goto failed_free;
@@ -264,20 +295,19 @@ static struct ip_tunnel * ipip_tunnel_locate(struct net *net,
        return nt;
 failed_free:
-        free_netdev(dev);
+        ipip_dev_free(dev);
        return NULL;
 }
+/* called with RTNL */
 static void ipip_tunnel_uninit(struct net_device *dev)
 {
        struct net *net = dev_net(dev);
        struct ipip_net *ipn = net_generic(net, ipip_net_id);
-        if (dev == ipn->fb_tunnel_dev) {
+        if (dev == ipn->fb_tunnel_dev)
-                spin_lock_bh(&ipip_lock);
+                rcu_assign_pointer(ipn->tunnels_wc[0], NULL);
-                ipn->tunnels_wc[0] = NULL;
+        else
-                spin_unlock_bh(&ipip_lock);
-        } else
                ipip_tunnel_unlink(ipn, netdev_priv(dev));
        dev_put(dev);
 }
@@ -359,8 +389,10 @@ static int ipip_rcv(struct sk_buff *skb)
        const struct iphdr *iph = ip_hdr(skb);
        rcu_read_lock();
-        if ((tunnel = ipip_tunnel_lookup(dev_net(skb->dev),
+        tunnel = ipip_tunnel_lookup(dev_net(skb->dev), iph->saddr, iph->daddr);
-                                        iph->saddr, iph->daddr)) != NULL) {
+        if (tunnel != NULL) {
+                struct pcpu_tstats *tstats;
                if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
                        rcu_read_unlock();
                        kfree_skb(skb);
@@ -374,10 +406,17 @@ static int ipip_rcv(struct sk_buff *skb)
                skb->protocol = htons(ETH_P_IP);
                skb->pkt_type = PACKET_HOST;
-                skb_tunnel_rx(skb, tunnel->dev);
+                tstats = this_cpu_ptr(tunnel->dev->tstats);
+                tstats->rx_packets++;
+                tstats->rx_bytes += skb->len;
+                __skb_tunnel_rx(skb, tunnel->dev);
                ipip_ecn_decapsulate(iph, skb);
-                netif_rx(skb);
+                if (netif_rx(skb) == NET_RX_DROP)
+                        tunnel->dev->stats.rx_dropped++;
                rcu_read_unlock();
                return 0;
        }
@@ -394,13 +433,12 @@ static int ipip_rcv(struct sk_buff *skb)
 static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
-        struct net_device_stats *stats = &dev->stats;
+        struct pcpu_tstats *tstats;
-        struct netdev_queue *txq = netdev_get_tx_queue(dev, 0);
        struct iphdr  *tiph = &tunnel->parms.iph;
        u8     tos = tunnel->parms.iph.tos;
        __be16 df = tiph->frag_off;
        struct rtable *rt;                      /* Route to the other host */
-        struct net_device *tdev;                        /* Device to other host */
+        struct net_device *tdev;                /* Device to other host */
        struct iphdr  *old_iph = ip_hdr(skb);
        struct iphdr  *iph;                     /* Our new IP header */
        unsigned int max_headroom;              /* The extra header space needed */
@@ -410,13 +448,13 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
        if (skb->protocol != htons(ETH_P_IP))
                goto tx_error;
-        if (tos&1)
+        if (tos & 1)
                tos = old_iph->tos;
        if (!dst) {
                /* NBMA tunnel */
                if ((rt = skb_rtable(skb)) == NULL) {
-                        stats->tx_fifo_errors++;
+                        dev->stats.tx_fifo_errors++;
                        goto tx_error;
                }
                if ((dst = rt->rt_gateway) == 0)
@@ -424,14 +462,20 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
        }
        {
-                struct flowi fl = { .oif = tunnel->parms.link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = tunnel->parms.link,
-                                              { .daddr = dst,
+                        .nl_u = {
-                                                .saddr = tiph->saddr,
+                                .ip4_u = {
-                                                .tos = RT_TOS(tos) } },
+                                        .daddr = dst,
-                                    .proto = IPPROTO_IPIP };
+                                        .saddr = tiph->saddr,
+                                        .tos = RT_TOS(tos)
+                                }
+                        },
+                        .proto = IPPROTO_IPIP
+                };
                if (ip_route_output_key(dev_net(dev), &rt, &fl)) {
-                        stats->tx_carrier_errors++;
+                        dev->stats.tx_carrier_errors++;
                        goto tx_error_icmp;
                }
        }
@@ -439,7 +483,7 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
        if (tdev == dev) {
                ip_rt_put(rt);
-                stats->collisions++;
+                dev->stats.collisions++;
                goto tx_error;
        }
@@ -449,7 +493,7 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
                mtu = dst_mtu(&rt->dst) - sizeof(struct iphdr);
                if (mtu < 68) {
-                        stats->collisions++;
+                        dev->stats.collisions++;
                        ip_rt_put(rt);
                        goto tx_error;
                }
@@ -485,7 +529,7 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
                struct sk_buff *new_skb = skb_realloc_headroom(skb, max_headroom);
                if (!new_skb) {
                        ip_rt_put(rt);
-                        txq->tx_dropped++;
+                        dev->stats.tx_dropped++;
                        dev_kfree_skb(skb);
                        return NETDEV_TX_OK;
                }
@@ -522,14 +566,14 @@ static netdev_tx_t ipip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev)
                iph->ttl        =       old_iph->ttl;
        nf_reset(skb);
+        tstats = this_cpu_ptr(dev->tstats);
-        IPTUNNEL_XMIT();
+        __IPTUNNEL_XMIT(tstats, &dev->stats);
        return NETDEV_TX_OK;
 tx_error_icmp:
        dst_link_failure(skb);
 tx_error:
-        stats->tx_errors++;
+        dev->stats.tx_errors++;
        dev_kfree_skb(skb);
        return NETDEV_TX_OK;
 }
@@ -544,13 +588,19 @@ static void ipip_tunnel_bind_dev(struct net_device *dev)
        iph = &tunnel->parms.iph;
        if (iph->daddr) {
-                struct flowi fl = { .oif = tunnel->parms.link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = tunnel->parms.link,
-                                              { .daddr = iph->daddr,
+                        .nl_u = {
-                                                .saddr = iph->saddr,
+                                .ip4_u = {
-                                                .tos = RT_TOS(iph->tos) } },
+                                        .daddr = iph->daddr,
-                                    .proto = IPPROTO_IPIP };
+                                        .saddr = iph->saddr,
+                                        .tos = RT_TOS(iph->tos)
+                                }
+                        },
+                        .proto = IPPROTO_IPIP
+                };
                struct rtable *rt;
                if (!ip_route_output_key(dev_net(dev), &rt, &fl)) {
                        tdev = rt->dst.dev;
                        ip_rt_put(rt);
@@ -696,13 +746,19 @@ static const struct net_device_ops ipip_netdev_ops = {
        .ndo_start_xmit = ipip_tunnel_xmit,
        .ndo_do_ioctl   = ipip_tunnel_ioctl,
        .ndo_change_mtu = ipip_tunnel_change_mtu,
+        .ndo_get_stats  = ipip_get_stats,
 };
+static void ipip_dev_free(struct net_device *dev)
+{
+        free_percpu(dev->tstats);
+        free_netdev(dev);
+}
 static void ipip_tunnel_setup(struct net_device *dev)
 {
        dev->netdev_ops         = &ipip_netdev_ops;
-        dev->destructor         = free_netdev;
+        dev->destructor         = ipip_dev_free;
        dev->type               = ARPHRD_TUNNEL;
        dev->hard_header_len    = LL_MAX_HEADER + sizeof(struct iphdr);
@@ -711,10 +767,11 @@ static void ipip_tunnel_setup(struct net_device *dev)
        dev->iflink             = 0;
        dev->addr_len           = 4;
        dev->features           |= NETIF_F_NETNS_LOCAL;
+        dev->features           |= NETIF_F_LLTX;
        dev->priv_flags         &= ~IFF_XMIT_DST_RELEASE;
 }
-static void ipip_tunnel_init(struct net_device *dev)
+static int ipip_tunnel_init(struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
@@ -725,9 +782,15 @@ static void ipip_tunnel_init(struct net_device *dev)
        memcpy(dev->broadcast, &tunnel->parms.iph.daddr, 4);
        ipip_tunnel_bind_dev(dev);
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
+        return 0;
 }
-static void __net_init ipip_fb_tunnel_init(struct net_device *dev)
+static int __net_init ipip_fb_tunnel_init(struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
        struct iphdr *iph = &tunnel->parms.iph;
@@ -740,11 +803,16 @@ static void __net_init ipip_fb_tunnel_init(struct net_device *dev)
        iph->protocol           = IPPROTO_IPIP;
        iph->ihl                = 5;
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
        dev_hold(dev);
-        ipn->tunnels_wc[0]      = tunnel;
+        rcu_assign_pointer(ipn->tunnels_wc[0], tunnel);
+        return 0;
 }
-static struct xfrm_tunnel ipip_handler = {
+static struct xfrm_tunnel ipip_handler __read_mostly = {
        .handler        =       ipip_rcv,
        .err_handler    =       ipip_err,
        .priority       =       1,
@@ -760,11 +828,12 @@ static void ipip_destroy_tunnels(struct ipip_net *ipn, struct list_head *head)
        for (prio = 1; prio < 4; prio++) {
                int h;
                for (h = 0; h < HASH_SIZE; h++) {
-                        struct ip_tunnel *t = ipn->tunnels[prio][h];
+                        struct ip_tunnel *t;
+                        t = rtnl_dereference(ipn->tunnels[prio][h]);
                        while (t != NULL) {
                                unregister_netdevice_queue(t->dev, head);
-                                t = t->next;
+                                t = rtnl_dereference(t->next);
                        }
                }
        }
@@ -789,7 +858,9 @@ static int __net_init ipip_init_net(struct net *net)
        }
        dev_net_set(ipn->fb_tunnel_dev, net);
-        ipip_fb_tunnel_init(ipn->fb_tunnel_dev);
+        err = ipip_fb_tunnel_init(ipn->fb_tunnel_dev);
+        if (err)
+                goto err_reg_dev;
        if ((err = register_netdev(ipn->fb_tunnel_dev)))
                goto err_reg_dev;
@@ -797,7 +868,7 @@ static int __net_init ipip_init_net(struct net *net)
        return 0;
 err_reg_dev:
-        free_netdev(ipn->fb_tunnel_dev);
+        ipip_dev_free(ipn->fb_tunnel_dev);
 err_alloc_dev:
        /* nothing */
        return err;
diff --git a/net/ipv4/ipmr.c b/net/ipv4/ipmr.c
index 179fcab866fc..86dd5691af46 100644
--- a/net/ipv4/ipmr.c
+++ b/net/ipv4/ipmr.c
@@ -75,7 +75,7 @@ struct mr_table {
        struct net              *net;
 #endif
        u32                     id;
-        struct sock             *mroute_sk;
+        struct sock __rcu       *mroute_sk;
        struct timer_list       ipmr_expire_timer;
        struct list_head        mfc_unres_queue;
        struct list_head        mfc_cache_array[MFC_LINES];
@@ -98,7 +98,7 @@ struct ipmr_result {
 };
 /* Big lock, protecting vif table, mrt cache and mroute socket state.
-   Note that the changes are semaphored via rtnl_lock.
+ * Note that the changes are semaphored via rtnl_lock.
 */
 static DEFINE_RWLOCK(mrt_lock);
@@ -113,11 +113,11 @@ static DEFINE_RWLOCK(mrt_lock);
 static DEFINE_SPINLOCK(mfc_unres_lock);
 /* We return to original Alan's scheme. Hash table of resolved
-   entries is changed only in process context and protected
+ * entries is changed only in process context and protected
-   with weak lock mrt_lock. Queue of unresolved entries is protected
+ * with weak lock mrt_lock. Queue of unresolved entries is protected
-   with strong spinlock mfc_unres_lock.
+ * with strong spinlock mfc_unres_lock.
+ *
-   In this case data path is free of exclusive locks at all.
+ * In this case data path is free of exclusive locks at all.
 */
 static struct kmem_cache *mrt_cachep __read_mostly;
@@ -396,9 +396,9 @@ struct net_device *ipmr_new_tunnel(struct net *net, struct vifctl *v)
                        set_fs(KERNEL_DS);
                        err = ops->ndo_do_ioctl(dev, &ifr, SIOCADDTUNNEL);
                        set_fs(oldfs);
-                } else
+                } else {
                        err = -EOPNOTSUPP;
+                }
                dev = NULL;
                if (err == 0 &&
@@ -495,7 +495,8 @@ static struct net_device *ipmr_reg_vif(struct net *net, struct mr_table *mrt)
        dev->iflink = 0;
        rcu_read_lock();
-        if ((in_dev = __in_dev_get_rcu(dev)) == NULL) {
+        in_dev = __in_dev_get_rcu(dev);
+        if (!in_dev) {
                rcu_read_unlock();
                goto failure;
        }
@@ -552,9 +553,10 @@ static int vif_delete(struct mr_table *mrt, int vifi, int notify,
                mrt->mroute_reg_vif_num = -1;
 #endif
-        if (vifi+1 == mrt->maxvif) {
+        if (vifi + 1 == mrt->maxvif) {
                int tmp;
-                for (tmp=vifi-1; tmp>=0; tmp--) {
+                for (tmp = vifi - 1; tmp >= 0; tmp--) {
                        if (VIF_EXISTS(mrt, tmp))
                                break;
                }
@@ -565,25 +567,33 @@ static int vif_delete(struct mr_table *mrt, int vifi, int notify,
        dev_set_allmulti(dev, -1);
-        if ((in_dev = __in_dev_get_rtnl(dev)) != NULL) {
+        in_dev = __in_dev_get_rtnl(dev);
+        if (in_dev) {
                IPV4_DEVCONF(in_dev->cnf, MC_FORWARDING)--;
                ip_rt_multicast_event(in_dev);
        }
-        if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER) && !notify)
+        if (v->flags & (VIFF_TUNNEL | VIFF_REGISTER) && !notify)
                unregister_netdevice_queue(dev, head);
        dev_put(dev);
        return 0;
 }
-static inline void ipmr_cache_free(struct mfc_cache *c)
+static void ipmr_cache_free_rcu(struct rcu_head *head)
 {
+        struct mfc_cache *c = container_of(head, struct mfc_cache, rcu);
        kmem_cache_free(mrt_cachep, c);
 }
+static inline void ipmr_cache_free(struct mfc_cache *c)
+{
+        call_rcu(&c->rcu, ipmr_cache_free_rcu);
+}
 /* Destroy an unresolved cache entry, killing queued skbs
-   and reporting error to netlink readers.
+ * and reporting error to netlink readers.
 */
 static void ipmr_destroy_unres(struct mr_table *mrt, struct mfc_cache *c)
@@ -605,8 +615,9 @@ static void ipmr_destroy_unres(struct mr_table *mrt, struct mfc_cache *c)
                        memset(&e->msg, 0, sizeof(e->msg));
                        rtnl_unicast(skb, net, NETLINK_CB(skb).pid);
-                } else
+                } else {
                        kfree_skb(skb);
+                }
        }
        ipmr_cache_free(c);
@@ -724,13 +735,13 @@ static int vif_add(struct net *net, struct mr_table *mrt,
        case 0:
                if (vifc->vifc_flags == VIFF_USE_IFINDEX) {
                        dev = dev_get_by_index(net, vifc->vifc_lcl_ifindex);
-                        if (dev && dev->ip_ptr == NULL) {
+                        if (dev && __in_dev_get_rtnl(dev) == NULL) {
                                dev_put(dev);
                                return -EADDRNOTAVAIL;
                        }
-                } else
+                } else {
                        dev = ip_dev_find(net, vifc->vifc_lcl_addr.s_addr);
+                }
                if (!dev)
                        return -EADDRNOTAVAIL;
                err = dev_set_allmulti(dev, 1);
@@ -743,16 +754,16 @@ static int vif_add(struct net *net, struct mr_table *mrt,
                return -EINVAL;
        }
-        if ((in_dev = __in_dev_get_rtnl(dev)) == NULL) {
+        in_dev = __in_dev_get_rtnl(dev);
+        if (!in_dev) {
                dev_put(dev);
                return -EADDRNOTAVAIL;
        }
        IPV4_DEVCONF(in_dev->cnf, MC_FORWARDING)++;
        ip_rt_multicast_event(in_dev);
-        /*
+        /* Fill in the VIF structures */
-         *      Fill in the VIF structures
-         */
        v->rate_limit = vifc->vifc_rate_limit;
        v->local = vifc->vifc_lcl_addr.s_addr;
        v->remote = vifc->vifc_rmt_addr.s_addr;
@@ -765,14 +776,14 @@ static int vif_add(struct net *net, struct mr_table *mrt,
        v->pkt_in = 0;
        v->pkt_out = 0;
        v->link = dev->ifindex;
-        if (v->flags&(VIFF_TUNNEL|VIFF_REGISTER))
+        if (v->flags & (VIFF_TUNNEL | VIFF_REGISTER))
                v->link = dev->iflink;
        /* And finish update writing critical data */
        write_lock_bh(&mrt_lock);
        v->dev = dev;
 #ifdef CONFIG_IP_PIMSM
-        if (v->flags&VIFF_REGISTER)
+        if (v->flags & VIFF_REGISTER)
                mrt->mroute_reg_vif_num = vifi;
 #endif
        if (vifi+1 > mrt->maxvif)
@@ -781,6 +792,7 @@ static int vif_add(struct net *net, struct mr_table *mrt,
        return 0;
 }
+/* called with rcu_read_lock() */
 static struct mfc_cache *ipmr_cache_find(struct mr_table *mrt,
                                         __be32 origin,
                                         __be32 mcastgrp)
@@ -788,7 +800,7 @@ static struct mfc_cache *ipmr_cache_find(struct mr_table *mrt,
        int line = MFC_HASH(mcastgrp, origin);
        struct mfc_cache *c;
-        list_for_each_entry(c, &mrt->mfc_cache_array[line], list) {
+        list_for_each_entry_rcu(c, &mrt->mfc_cache_array[line], list) {
                if (c->mfc_origin == origin && c->mfc_mcastgrp == mcastgrp)
                        return c;
        }
@@ -801,19 +813,20 @@ static struct mfc_cache *ipmr_cache_find(struct mr_table *mrt,
 static struct mfc_cache *ipmr_cache_alloc(void)
 {
        struct mfc_cache *c = kmem_cache_zalloc(mrt_cachep, GFP_KERNEL);
-        if (c == NULL)
-                return NULL;
+        if (c)
-        c->mfc_un.res.minvif = MAXVIFS;
+                c->mfc_un.res.minvif = MAXVIFS;
        return c;
 }
 static struct mfc_cache *ipmr_cache_alloc_unres(void)
 {
        struct mfc_cache *c = kmem_cache_zalloc(mrt_cachep, GFP_ATOMIC);
-        if (c == NULL)
-                return NULL;
+        if (c) {
-        skb_queue_head_init(&c->mfc_un.unres.unresolved);
+                skb_queue_head_init(&c->mfc_un.unres.unresolved);
-        c->mfc_un.unres.expires = jiffies + 10*HZ;
+                c->mfc_un.unres.expires = jiffies + 10*HZ;
+        }
        return c;
 }
@@ -827,17 +840,15 @@ static void ipmr_cache_resolve(struct net *net, struct mr_table *mrt,
        struct sk_buff *skb;
        struct nlmsgerr *e;
-        /*
+        /* Play the pending entries through our router */
-         *      Play the pending entries through our router
-         */
        while ((skb = __skb_dequeue(&uc->mfc_un.unres.unresolved))) {
                if (ip_hdr(skb)->version == 0) {
                        struct nlmsghdr *nlh = (struct nlmsghdr *)skb_pull(skb, sizeof(struct iphdr));
                        if (__ipmr_fill_mroute(mrt, skb, c, NLMSG_DATA(nlh)) > 0) {
-                                nlh->nlmsg_len = (skb_tail_pointer(skb) -
+                                nlh->nlmsg_len = skb_tail_pointer(skb) -
-                                                  (u8 *)nlh);
+                                                 (u8 *)nlh;
                        } else {
                                nlh->nlmsg_type = NLMSG_ERROR;
                                nlh->nlmsg_len = NLMSG_LENGTH(sizeof(struct nlmsgerr));
@@ -848,8 +859,9 @@ static void ipmr_cache_resolve(struct net *net, struct mr_table *mrt,
                        }
                        rtnl_unicast(skb, net, NETLINK_CB(skb).pid);
-                } else
+                } else {
                        ip_mr_forward(net, mrt, skb, c, 0);
+                }
        }
 }
@@ -867,6 +879,7 @@ static int ipmr_cache_report(struct mr_table *mrt,
        const int ihl = ip_hdrlen(pkt);
        struct igmphdr *igmp;
        struct igmpmsg *msg;
+        struct sock *mroute_sk;
        int ret;
 #ifdef CONFIG_IP_PIMSM
@@ -882,9 +895,9 @@ static int ipmr_cache_report(struct mr_table *mrt,
 #ifdef CONFIG_IP_PIMSM
        if (assert == IGMPMSG_WHOLEPKT) {
                /* Ugly, but we have no choice with this interface.
-                   Duplicate old header, fix ihl, length etc.
+                 * Duplicate old header, fix ihl, length etc.
-                   And all this only to mangle msg->im_msgtype and
+                 * And all this only to mangle msg->im_msgtype and
-                   to set msg->im_mbz to "mbz" :-)
+                 * to set msg->im_mbz to "mbz" :-)
                 */
                skb_push(skb, sizeof(struct iphdr));
                skb_reset_network_header(skb);
@@ -901,39 +914,38 @@ static int ipmr_cache_report(struct mr_table *mrt,
 #endif
        {
-        /*
+        /* Copy the IP header */
-         *      Copy the IP header
-         */
        skb->network_header = skb->tail;
        skb_put(skb, ihl);
        skb_copy_to_linear_data(skb, pkt->data, ihl);
-        ip_hdr(skb)->protocol = 0;                      /* Flag to the kernel this is a route add */
+        ip_hdr(skb)->protocol = 0;      /* Flag to the kernel this is a route add */
        msg = (struct igmpmsg *)skb_network_header(skb);
        msg->im_vif = vifi;
        skb_dst_set(skb, dst_clone(skb_dst(pkt)));
-        /*
+        /* Add our header */
-         *      Add our header
-         */
-        igmp=(struct igmphdr *)skb_put(skb, sizeof(struct igmphdr));
+        igmp = (struct igmphdr *)skb_put(skb, sizeof(struct igmphdr));
        igmp->type      =
        msg->im_msgtype = assert;
-        igmp->code      =       0;
+        igmp->code      = 0;
-        ip_hdr(skb)->tot_len = htons(skb->len);                 /* Fix the length */
+        ip_hdr(skb)->tot_len = htons(skb->len);         /* Fix the length */
        skb->transport_header = skb->network_header;
        }
-        if (mrt->mroute_sk == NULL) {
+        rcu_read_lock();
+        mroute_sk = rcu_dereference(mrt->mroute_sk);
+        if (mroute_sk == NULL) {
+                rcu_read_unlock();
                kfree_skb(skb);
                return -EINVAL;
        }
-        /*
+        /* Deliver to mrouted */
-         *      Deliver to mrouted
-         */
+        ret = sock_queue_rcv_skb(mroute_sk, skb);
-        ret = sock_queue_rcv_skb(mrt->mroute_sk, skb);
+        rcu_read_unlock();
        if (ret < 0) {
                if (net_ratelimit())
                        printk(KERN_WARNING "mroute: pending queue full, dropping entries.\n");
@@ -965,9 +977,7 @@ ipmr_cache_unresolved(struct mr_table *mrt, vifi_t vifi, struct sk_buff *skb)
        }
        if (!found) {
-                /*
+                /* Create a new entry if allowable */
-                 *      Create a new entry if allowable
-                 */
                if (atomic_read(&mrt->cache_resolve_queue_len) >= 10 ||
                    (c = ipmr_cache_alloc_unres()) == NULL) {
@@ -977,16 +987,14 @@ ipmr_cache_unresolved(struct mr_table *mrt, vifi_t vifi, struct sk_buff *skb)
                        return -ENOBUFS;
                }
-                /*
+                /* Fill in the new cache entry */
-                 *      Fill in the new cache entry
-                 */
                c->mfc_parent   = -1;
                c->mfc_origin   = iph->saddr;
                c->mfc_mcastgrp = iph->daddr;
-                /*
+                /* Reflect first query at mrouted. */
-                 *      Reflect first query at mrouted.
-                 */
                err = ipmr_cache_report(mrt, skb, vifi, IGMPMSG_NOCACHE);
                if (err < 0) {
                        /* If the report failed throw the cache entry
@@ -1006,10 +1014,9 @@ ipmr_cache_unresolved(struct mr_table *mrt, vifi_t vifi, struct sk_buff *skb)
                        mod_timer(&mrt->ipmr_expire_timer, c->mfc_un.unres.expires);
        }
-        /*
+        /* See if we can append the packet */
-         *      See if we can append the packet
-         */
+        if (c->mfc_un.unres.unresolved.qlen > 3) {
-        if (c->mfc_un.unres.unresolved.qlen>3) {
                kfree_skb(skb);
                err = -ENOBUFS;
        } else {
@@ -1035,9 +1042,7 @@ static int ipmr_mfc_delete(struct mr_table *mrt, struct mfcctl *mfc)
        list_for_each_entry_safe(c, next, &mrt->mfc_cache_array[line], list) {
                if (c->mfc_origin == mfc->mfcc_origin.s_addr &&
                    c->mfc_mcastgrp == mfc->mfcc_mcastgrp.s_addr) {
-                        write_lock_bh(&mrt_lock);
+                        list_del_rcu(&c->list);
-                        list_del(&c->list);
-                        write_unlock_bh(&mrt_lock);
                        ipmr_cache_free(c);
                        return 0;
@@ -1090,9 +1095,7 @@ static int ipmr_mfc_add(struct net *net, struct mr_table *mrt,
        if (!mrtsock)
                c->mfc_flags |= MFC_STATIC;
-        write_lock_bh(&mrt_lock);
+        list_add_rcu(&c->list, &mrt->mfc_cache_array[line]);
-        list_add(&c->list, &mrt->mfc_cache_array[line]);
-        write_unlock_bh(&mrt_lock);
        /*
         *      Check to see if we resolved a queued list. If so we
@@ -1130,26 +1133,21 @@ static void mroute_clean_tables(struct mr_table *mrt)
        LIST_HEAD(list);
        struct mfc_cache *c, *next;
-        /*
+        /* Shut down all active vif entries */
-         *      Shut down all active vif entries
-         */
        for (i = 0; i < mrt->maxvif; i++) {
-                if (!(mrt->vif_table[i].flags&VIFF_STATIC))
+                if (!(mrt->vif_table[i].flags & VIFF_STATIC))
                        vif_delete(mrt, i, 0, &list);
        }
        unregister_netdevice_many(&list);
-        /*
+        /* Wipe the cache */
-         *      Wipe the cache
-         */
        for (i = 0; i < MFC_LINES; i++) {
                list_for_each_entry_safe(c, next, &mrt->mfc_cache_array[i], list) {
-                        if (c->mfc_flags&MFC_STATIC)
+                        if (c->mfc_flags & MFC_STATIC)
                                continue;
-                        write_lock_bh(&mrt_lock);
+                        list_del_rcu(&c->list);
-                        list_del(&c->list);
-                        write_unlock_bh(&mrt_lock);
                        ipmr_cache_free(c);
                }
        }
@@ -1164,6 +1162,9 @@ static void mroute_clean_tables(struct mr_table *mrt)
        }
 }
+/* called from ip_ra_control(), before an RCU grace period,
+ * we dont need to call synchronize_rcu() here
+ */
 static void mrtsock_destruct(struct sock *sk)
 {
        struct net *net = sock_net(sk);
@@ -1171,13 +1172,9 @@ static void mrtsock_destruct(struct sock *sk)
        rtnl_lock();
        ipmr_for_each_table(mrt, net) {
-                if (sk == mrt->mroute_sk) {
+                if (sk == rtnl_dereference(mrt->mroute_sk)) {
                        IPV4_DEVCONF_ALL(net, MC_FORWARDING)--;
+                        rcu_assign_pointer(mrt->mroute_sk, NULL);
-                        write_lock_bh(&mrt_lock);
-                        mrt->mroute_sk = NULL;
-                        write_unlock_bh(&mrt_lock);
                        mroute_clean_tables(mrt);
                }
        }
@@ -1204,7 +1201,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
                return -ENOENT;
        if (optname != MRT_INIT) {
-                if (sk != mrt->mroute_sk && !capable(CAP_NET_ADMIN))
+                if (sk != rcu_dereference_raw(mrt->mroute_sk) &&
+                    !capable(CAP_NET_ADMIN))
                        return -EACCES;
        }
@@ -1217,23 +1215,20 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
                        return -ENOPROTOOPT;
                rtnl_lock();
-                if (mrt->mroute_sk) {
+                if (rtnl_dereference(mrt->mroute_sk)) {
                        rtnl_unlock();
                        return -EADDRINUSE;
                }
                ret = ip_ra_control(sk, 1, mrtsock_destruct);
                if (ret == 0) {
-                        write_lock_bh(&mrt_lock);
+                        rcu_assign_pointer(mrt->mroute_sk, sk);
-                        mrt->mroute_sk = sk;
-                        write_unlock_bh(&mrt_lock);
                        IPV4_DEVCONF_ALL(net, MC_FORWARDING)++;
                }
                rtnl_unlock();
                return ret;
        case MRT_DONE:
-                if (sk != mrt->mroute_sk)
+                if (sk != rcu_dereference_raw(mrt->mroute_sk))
                        return -EACCES;
                return ip_ra_control(sk, 0, NULL);
        case MRT_ADD_VIF:
@@ -1246,7 +1241,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
                        return -ENFILE;
                rtnl_lock();
                if (optname == MRT_ADD_VIF) {
-                        ret = vif_add(net, mrt, &vif, sk == mrt->mroute_sk);
+                        ret = vif_add(net, mrt, &vif,
+                                      sk == rtnl_dereference(mrt->mroute_sk));
                } else {
                        ret = vif_delete(mrt, vif.vifc_vifi, 0, NULL);
                }
@@ -1267,7 +1263,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
                if (optname == MRT_DEL_MFC)
                        ret = ipmr_mfc_delete(mrt, &mfc);
                else
-                        ret = ipmr_mfc_add(net, mrt, &mfc, sk == mrt->mroute_sk);
+                        ret = ipmr_mfc_add(net, mrt, &mfc,
+                                           sk == rtnl_dereference(mrt->mroute_sk));
                rtnl_unlock();
                return ret;
                /*
@@ -1276,7 +1273,7 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
        case MRT_ASSERT:
        {
                int v;
-                if (get_user(v,(int __user *)optval))
+                if (get_user(v, (int __user *)optval))
                        return -EFAULT;
                mrt->mroute_do_assert = (v) ? 1 : 0;
                return 0;
@@ -1286,7 +1283,7 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
        {
                int v;
-                if (get_user(v,(int __user *)optval))
+                if (get_user(v, (int __user *)optval))
                        return -EFAULT;
                v = (v) ? 1 : 0;
@@ -1309,14 +1306,16 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
                        return -EINVAL;
                if (get_user(v, (u32 __user *)optval))
                        return -EFAULT;
-                if (sk == mrt->mroute_sk)
-                        return -EBUSY;
                rtnl_lock();
                ret = 0;
-                if (!ipmr_new_table(net, v))
+                if (sk == rtnl_dereference(mrt->mroute_sk)) {
-                        ret = -ENOMEM;
+                        ret = -EBUSY;
-                raw_sk(sk)->ipmr_table = v;
+                } else {
+                        if (!ipmr_new_table(net, v))
+                                ret = -ENOMEM;
+                        raw_sk(sk)->ipmr_table = v;
+                }
                rtnl_unlock();
                return ret;
        }
@@ -1347,9 +1346,9 @@ int ip_mroute_getsockopt(struct sock *sk, int optname, char __user *optval, int
        if (optname != MRT_VERSION &&
 #ifdef CONFIG_IP_PIMSM
-           optname!=MRT_PIM &&
+           optname != MRT_PIM &&
 #endif
-           optname!=MRT_ASSERT)
+           optname != MRT_ASSERT)
                return -ENOPROTOOPT;
        if (get_user(olr, optlen))
@@ -1416,19 +1415,19 @@ int ipmr_ioctl(struct sock *sk, int cmd, void __user *arg)
                if (copy_from_user(&sr, arg, sizeof(sr)))
                        return -EFAULT;
-                read_lock(&mrt_lock);
+                rcu_read_lock();
                c = ipmr_cache_find(mrt, sr.src.s_addr, sr.grp.s_addr);
                if (c) {
                        sr.pktcnt = c->mfc_un.res.pkt;
                        sr.bytecnt = c->mfc_un.res.bytes;
                        sr.wrong_if = c->mfc_un.res.wrong_if;
-                        read_unlock(&mrt_lock);
+                        rcu_read_unlock();
                        if (copy_to_user(arg, &sr, sizeof(sr)))
                                return -EFAULT;
                        return 0;
                }
-                read_unlock(&mrt_lock);
+                rcu_read_unlock();
                return -EADDRNOTAVAIL;
        default:
                return -ENOIOCTLCMD;
@@ -1465,7 +1464,7 @@ static struct notifier_block ip_mr_notifier = {
 };
 /*
- *      Encapsulate a packet by attaching a valid IPIP header to it.
+ *      Encapsulate a packet by attaching a valid IPIP header to it.
 *      This avoids tunnel drivers and other mess and gives us the speed so
 *      important for multicast video.
 */
@@ -1480,7 +1479,7 @@ static void ip_encap(struct sk_buff *skb, __be32 saddr, __be32 daddr)
        skb_reset_network_header(skb);
        iph = ip_hdr(skb);
-        iph->version    =       4;
+        iph->version    =       4;
        iph->tos        =       old_iph->tos;
        iph->ttl        =       old_iph->ttl;
        iph->frag_off   =       0;
@@ -1498,7 +1497,7 @@ static void ip_encap(struct sk_buff *skb, __be32 saddr, __be32 daddr)
 static inline int ipmr_forward_finish(struct sk_buff *skb)
 {
-        struct ip_options * opt = &(IPCB(skb)->opt);
+        struct ip_options *opt = &(IPCB(skb)->opt);
        IP_INC_STATS_BH(dev_net(skb_dst(skb)->dev), IPSTATS_MIB_OUTFORWDATAGRAMS);
@@ -1535,22 +1534,34 @@ static void ipmr_queue_xmit(struct net *net, struct mr_table *mrt,
        }
 #endif
-        if (vif->flags&VIFF_TUNNEL) {
+        if (vif->flags & VIFF_TUNNEL) {
-                struct flowi fl = { .oif = vif->link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = vif->link,
-                                              { .daddr = vif->remote,
+                        .nl_u = {
-                                                .saddr = vif->local,
+                                .ip4_u = {
-                                                .tos = RT_TOS(iph->tos) } },
+                                        .daddr = vif->remote,
-                                    .proto = IPPROTO_IPIP };
+                                        .saddr = vif->local,
+                                        .tos = RT_TOS(iph->tos)
+                                }
+                        },
+                        .proto = IPPROTO_IPIP
+                };
                if (ip_route_output_key(net, &rt, &fl))
                        goto out_free;
                encap = sizeof(struct iphdr);
        } else {
-                struct flowi fl = { .oif = vif->link,
+                struct flowi fl = {
-                                    .nl_u = { .ip4_u =
+                        .oif = vif->link,
-                                              { .daddr = iph->daddr,
+                        .nl_u = {
-                                                .tos = RT_TOS(iph->tos) } },
+                                .ip4_u = {
-                                    .proto = IPPROTO_IPIP };
+                                        .daddr = iph->daddr,
+                                        .tos = RT_TOS(iph->tos)
+                                }
+                        },
+                        .proto = IPPROTO_IPIP
+                };
                if (ip_route_output_key(net, &rt, &fl))
                        goto out_free;
        }
@@ -1559,8 +1570,8 @@ static void ipmr_queue_xmit(struct net *net, struct mr_table *mrt,
        if (skb->len+encap > dst_mtu(&rt->dst) && (ntohs(iph->frag_off) & IP_DF)) {
                /* Do not fragment multicasts. Alas, IPv4 does not
-                   allow to send ICMP, so that packets will disappear
+                 * allow to send ICMP, so that packets will disappear
-                   to blackhole.
+                 * to blackhole.
                 */
                IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_FRAGFAILS);
@@ -1583,7 +1594,8 @@ static void ipmr_queue_xmit(struct net *net, struct mr_table *mrt,
        ip_decrease_ttl(ip_hdr(skb));
        /* FIXME: forward and output firewalls used to be called here.
-         * What do we do with netfilter? -- RR */
+         * What do we do with netfilter? -- RR
+         */
        if (vif->flags & VIFF_TUNNEL) {
                ip_encap(skb, vif->local, vif->remote);
                /* FIXME: extra output firewall step used to be here. --RR */
@@ -1644,15 +1656,15 @@ static int ip_mr_forward(struct net *net, struct mr_table *mrt,
                if (skb_rtable(skb)->fl.iif == 0) {
                        /* It is our own packet, looped back.
-                           Very complicated situation...
+                         * Very complicated situation...
+                         *
-                           The best workaround until routing daemons will be
+                         * The best workaround until routing daemons will be
-                           fixed is not to redistribute packet, if it was
+                         * fixed is not to redistribute packet, if it was
-                           send through wrong interface. It means, that
+                         * send through wrong interface. It means, that
-                           multicast applications WILL NOT work for
+                         * multicast applications WILL NOT work for
-                           (S,G), which have default multicast route pointing
+                         * (S,G), which have default multicast route pointing
-                           to wrong oif. In any case, it is not a good
+                         * to wrong oif. In any case, it is not a good
-                           idea to use multicasting applications on router.
+                         * idea to use multicasting applications on router.
                         */
                        goto dont_forward;
                }
@@ -1662,9 +1674,9 @@ static int ip_mr_forward(struct net *net, struct mr_table *mrt,
                if (true_vifi >= 0 && mrt->mroute_do_assert &&
                    /* pimsm uses asserts, when switching from RPT to SPT,
-                       so that we cannot check that packet arrived on an oif.
+                     * so that we cannot check that packet arrived on an oif.
-                       It is bad, but otherwise we would need to move pretty
+                     * It is bad, but otherwise we would need to move pretty
-                       large chunk of pimd to kernel. Ough... --ANK
+                     * large chunk of pimd to kernel. Ough... --ANK
                     */
                    (mrt->mroute_do_pim ||
                     cache->mfc_un.res.ttls[true_vifi] < 255) &&
@@ -1682,10 +1694,12 @@ static int ip_mr_forward(struct net *net, struct mr_table *mrt,
        /*
         *      Forward the frame
         */
-        for (ct = cache->mfc_un.res.maxvif-1; ct >= cache->mfc_un.res.minvif; ct--) {
+        for (ct = cache->mfc_un.res.maxvif - 1;
+             ct >= cache->mfc_un.res.minvif; ct--) {
                if (ip_hdr(skb)->ttl > cache->mfc_un.res.ttls[ct]) {
                        if (psend != -1) {
                                struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
                                if (skb2)
                                        ipmr_queue_xmit(net, mrt, skb2, cache,
                                                        psend);
@@ -1696,6 +1710,7 @@ static int ip_mr_forward(struct net *net, struct mr_table *mrt,
        if (psend != -1) {
                if (local) {
                        struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
                        if (skb2)
                                ipmr_queue_xmit(net, mrt, skb2, cache, psend);
                } else {
@@ -1713,6 +1728,7 @@ dont_forward:
 /*
 *      Multicast packets for forwarding arrive here
+ *      Called with rcu_read_lock();
 */
 int ip_mr_input(struct sk_buff *skb)
@@ -1724,9 +1740,9 @@ int ip_mr_input(struct sk_buff *skb)
        int err;
        /* Packet is looped back after forward, it should not be
-           forwarded second time, but still can be delivered locally.
+         * forwarded second time, but still can be delivered locally.
         */
-        if (IPCB(skb)->flags&IPSKB_FORWARDED)
+        if (IPCB(skb)->flags & IPSKB_FORWARDED)
                goto dont_forward;
        err = ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt);
@@ -1736,28 +1752,28 @@ int ip_mr_input(struct sk_buff *skb)
        }
        if (!local) {
-                    if (IPCB(skb)->opt.router_alert) {
+                if (IPCB(skb)->opt.router_alert) {
-                            if (ip_call_ra_chain(skb))
+                        if (ip_call_ra_chain(skb))
-                                    return 0;
+                                return 0;
-                    } else if (ip_hdr(skb)->protocol == IPPROTO_IGMP){
+                } else if (ip_hdr(skb)->protocol == IPPROTO_IGMP) {
-                            /* IGMPv1 (and broken IGMPv2 implementations sort of
+                        /* IGMPv1 (and broken IGMPv2 implementations sort of
-                               Cisco IOS <= 11.2(8)) do not put router alert
+                         * Cisco IOS <= 11.2(8)) do not put router alert
-                               option to IGMP packets destined to routable
+                         * option to IGMP packets destined to routable
-                               groups. It is very bad, because it means
+                         * groups. It is very bad, because it means
-                               that we can forward NO IGMP messages.
+                         * that we can forward NO IGMP messages.
-                             */
+                         */
-                            read_lock(&mrt_lock);
+                        struct sock *mroute_sk;
-                            if (mrt->mroute_sk) {
-                                    nf_reset(skb);
+                        mroute_sk = rcu_dereference(mrt->mroute_sk);
-                                    raw_rcv(mrt->mroute_sk, skb);
+                        if (mroute_sk) {
-                                    read_unlock(&mrt_lock);
+                                nf_reset(skb);
-                                    return 0;
+                                raw_rcv(mroute_sk, skb);
-                            }
+                                return 0;
-                            read_unlock(&mrt_lock);
+                        }
                    }
        }
-        read_lock(&mrt_lock);
+        /* already under rcu_read_lock() */
        cache = ipmr_cache_find(mrt, ip_hdr(skb)->saddr, ip_hdr(skb)->daddr);
        /*
@@ -1769,13 +1785,12 @@ int ip_mr_input(struct sk_buff *skb)
                if (local) {
                        struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC);
                        ip_local_deliver(skb);
-                        if (skb2 == NULL) {
+                        if (skb2 == NULL)
-                                read_unlock(&mrt_lock);
                                return -ENOBUFS;
-                        }
                        skb = skb2;
                }
+                read_lock(&mrt_lock);
                vif = ipmr_find_vif(mrt, skb->dev);
                if (vif >= 0) {
                        int err2 = ipmr_cache_unresolved(mrt, vif, skb);
@@ -1788,8 +1803,8 @@ int ip_mr_input(struct sk_buff *skb)
                return -ENODEV;
        }
+        read_lock(&mrt_lock);
        ip_mr_forward(net, mrt, skb, cache, local);
        read_unlock(&mrt_lock);
        if (local)
@@ -1805,6 +1820,7 @@ dont_forward:
 }
 #ifdef CONFIG_IP_PIMSM
+/* called with rcu_read_lock() */
 static int __pim_rcv(struct mr_table *mrt, struct sk_buff *skb,
                     unsigned int pimlen)
 {
@@ -1813,10 +1829,10 @@ static int __pim_rcv(struct mr_table *mrt, struct sk_buff *skb,
        encap = (struct iphdr *)(skb_transport_header(skb) + pimlen);
        /*
-           Check that:
+         * Check that:
-           a. packet is really destinted to a multicast group
+         * a. packet is really sent to a multicast group
-           b. packet is not a NULL-REGISTER
+         * b. packet is not a NULL-REGISTER
-           c. packet is not truncated
+         * c. packet is not truncated
         */
        if (!ipv4_is_multicast(encap->daddr) ||
            encap->tot_len == 0 ||
@@ -1826,26 +1842,23 @@ static int __pim_rcv(struct mr_table *mrt, struct sk_buff *skb,
        read_lock(&mrt_lock);
        if (mrt->mroute_reg_vif_num >= 0)
                reg_dev = mrt->vif_table[mrt->mroute_reg_vif_num].dev;
-        if (reg_dev)
-                dev_hold(reg_dev);
        read_unlock(&mrt_lock);
        if (reg_dev == NULL)
                return 1;
        skb->mac_header = skb->network_header;
-        skb_pull(skb, (u8*)encap - skb->data);
+        skb_pull(skb, (u8 *)encap - skb->data);
        skb_reset_network_header(skb);
        skb->protocol = htons(ETH_P_IP);
-        skb->ip_summed = 0;
+        skb->ip_summed = CHECKSUM_NONE;
        skb->pkt_type = PACKET_HOST;
        skb_tunnel_rx(skb, reg_dev);
        netif_rx(skb);
-        dev_put(reg_dev);
-        return 0;
+        return NET_RX_SUCCESS;
 }
 #endif
@@ -1854,7 +1867,7 @@ static int __pim_rcv(struct mr_table *mrt, struct sk_buff *skb,
 * Handle IGMP messages of PIMv1
 */
-int pim_rcv_v1(struct sk_buff * skb)
+int pim_rcv_v1(struct sk_buff *skb)
 {
        struct igmphdr *pim;
        struct net *net = dev_net(skb->dev);
@@ -1881,7 +1894,7 @@ drop:
 #endif
 #ifdef CONFIG_IP_PIMSM_V2
-static int pim_rcv(struct sk_buff * skb)
+static int pim_rcv(struct sk_buff *skb)
 {
        struct pimreghdr *pim;
        struct net *net = dev_net(skb->dev);
@@ -1891,8 +1904,8 @@ static int pim_rcv(struct sk_buff * skb)
                goto drop;
        pim = (struct pimreghdr *)skb_transport_header(skb);
-        if (pim->type != ((PIM_VERSION<<4)|(PIM_REGISTER)) ||
+        if (pim->type != ((PIM_VERSION << 4) | (PIM_REGISTER)) ||
-            (pim->flags&PIM_NULL_REGISTER) ||
+            (pim->flags & PIM_NULL_REGISTER) ||
            (ip_compute_csum((void *)pim, sizeof(*pim)) != 0 &&
             csum_fold(skb_checksum(skb, 0, skb->len, 0))))
                goto drop;
@@ -1958,28 +1971,33 @@ int ipmr_get_route(struct net *net,
        if (mrt == NULL)
                return -ENOENT;
-        read_lock(&mrt_lock);
+        rcu_read_lock();
        cache = ipmr_cache_find(mrt, rt->rt_src, rt->rt_dst);
        if (cache == NULL) {
                struct sk_buff *skb2;
                struct iphdr *iph;
                struct net_device *dev;
-                int vif;
+                int vif = -1;
                if (nowait) {
-                        read_unlock(&mrt_lock);
+                        rcu_read_unlock();
                        return -EAGAIN;
                }
                dev = skb->dev;
-                if (dev == NULL || (vif = ipmr_find_vif(mrt, dev)) < 0) {
+                read_lock(&mrt_lock);
+                if (dev)
+                        vif = ipmr_find_vif(mrt, dev);
+                if (vif < 0) {
                        read_unlock(&mrt_lock);
+                        rcu_read_unlock();
                        return -ENODEV;
                }
                skb2 = skb_clone(skb, GFP_ATOMIC);
                if (!skb2) {
                        read_unlock(&mrt_lock);
+                        rcu_read_unlock();
                        return -ENOMEM;
                }
@@ -1992,13 +2010,16 @@ int ipmr_get_route(struct net *net,
                iph->version = 0;
                err = ipmr_cache_unresolved(mrt, vif, skb2);
                read_unlock(&mrt_lock);
+                rcu_read_unlock();
                return err;
        }
-        if (!nowait && (rtm->rtm_flags&RTM_F_NOTIFY))
+        read_lock(&mrt_lock);
+        if (!nowait && (rtm->rtm_flags & RTM_F_NOTIFY))
                cache->mfc_flags |= MFC_NOTIFY;
        err = __ipmr_fill_mroute(mrt, skb, cache, rtm);
        read_unlock(&mrt_lock);
+        rcu_read_unlock();
        return err;
 }
@@ -2050,14 +2071,14 @@ static int ipmr_rtm_dumproute(struct sk_buff *skb, struct netlink_callback *cb)
        s_h = cb->args[1];
        s_e = cb->args[2];
-        read_lock(&mrt_lock);
+        rcu_read_lock();
        ipmr_for_each_table(mrt, net) {
                if (t < s_t)
                        goto next_table;
                if (t > s_t)
                        s_h = 0;
                for (h = s_h; h < MFC_LINES; h++) {
-                        list_for_each_entry(mfc, &mrt->mfc_cache_array[h], list) {
+                        list_for_each_entry_rcu(mfc, &mrt->mfc_cache_array[h], list) {
                                if (e < s_e)
                                        goto next_entry;
                                if (ipmr_fill_mroute(mrt, skb,
@@ -2075,7 +2096,7 @@ next_table:
                t++;
        }
 done:
-        read_unlock(&mrt_lock);
+        rcu_read_unlock();
        cb->args[2] = e;
        cb->args[1] = h;
@@ -2086,7 +2107,8 @@ done:
 #ifdef CONFIG_PROC_FS
 /*
- *      The /proc interfaces to multicast routing /proc/ip_mr_cache /proc/ip_mr_vif
+ *      The /proc interfaces to multicast routing :
+ *      /proc/net/ip_mr_cache & /proc/net/ip_mr_vif
 */
 struct ipmr_vif_iter {
        struct seq_net_private p;
@@ -2208,14 +2230,14 @@ static struct mfc_cache *ipmr_mfc_seq_idx(struct net *net,
        struct mr_table *mrt = it->mrt;
        struct mfc_cache *mfc;
-        read_lock(&mrt_lock);
+        rcu_read_lock();
        for (it->ct = 0; it->ct < MFC_LINES; it->ct++) {
                it->cache = &mrt->mfc_cache_array[it->ct];
-                list_for_each_entry(mfc, it->cache, list)
+                list_for_each_entry_rcu(mfc, it->cache, list)
                        if (pos-- == 0)
                                return mfc;
        }
-        read_unlock(&mrt_lock);
+        rcu_read_unlock();
        spin_lock_bh(&mfc_unres_lock);
        it->cache = &mrt->mfc_unres_queue;
@@ -2274,7 +2296,7 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
        }
        /* exhausted cache_array, show unresolved */
-        read_unlock(&mrt_lock);
+        rcu_read_unlock();
        it->cache = &mrt->mfc_unres_queue;
        it->ct = 0;
@@ -2282,7 +2304,7 @@ static void *ipmr_mfc_seq_next(struct seq_file *seq, void *v, loff_t *pos)
        if (!list_empty(it->cache))
                return list_first_entry(it->cache, struct mfc_cache, list);
- end_of_list:
+end_of_list:
        spin_unlock_bh(&mfc_unres_lock);
        it->cache = NULL;
@@ -2297,7 +2319,7 @@ static void ipmr_mfc_seq_stop(struct seq_file *seq, void *v)
        if (it->cache == &mrt->mfc_unres_queue)
                spin_unlock_bh(&mfc_unres_lock);
        else if (it->cache == &mrt->mfc_cache_array[it->ct])
-                read_unlock(&mrt_lock);
+                rcu_read_unlock();
 }
 static int ipmr_mfc_seq_show(struct seq_file *seq, void *v)
@@ -2323,7 +2345,7 @@ static int ipmr_mfc_seq_show(struct seq_file *seq, void *v)
                                   mfc->mfc_un.res.bytes,
                                   mfc->mfc_un.res.wrong_if);
                        for (n = mfc->mfc_un.res.minvif;
-                             n < mfc->mfc_un.res.maxvif; n++ ) {
+                             n < mfc->mfc_un.res.maxvif; n++) {
                                if (VIF_EXISTS(mrt, n) &&
                                    mfc->mfc_un.res.ttls[n] < 255)
                                        seq_printf(seq,
@@ -2421,7 +2443,7 @@ int __init ip_mr_init(void)
        mrt_cachep = kmem_cache_create("ip_mrt_cache",
                                       sizeof(struct mfc_cache),
-                                       0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
+                                       0, SLAB_HWCACHE_ALIGN | SLAB_PANIC,
                                       NULL);
        if (!mrt_cachep)
                return -ENOMEM;
diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c
index e8f4f9a57f12..8b642f152468 100644
--- a/net/ipv4/netfilter/arp_tables.c
+++ b/net/ipv4/netfilter/arp_tables.c
@@ -72,7 +72,7 @@ static inline int arp_devaddr_compare(const struct arpt_devaddr_info *ap,
        for (i = 0; i < len; i++)
                ret |= (hdr_addr[i] ^ ap->addr[i]) & ap->mask[i];
-        return (ret != 0);
+        return ret != 0;
 }
 /*
diff --git a/net/ipv4/netfilter/ipt_CLUSTERIP.c b/net/ipv4/netfilter/ipt_CLUSTERIP.c
index 3a43cf36db87..1e26a4897655 100644
--- a/net/ipv4/netfilter/ipt_CLUSTERIP.c
+++ b/net/ipv4/netfilter/ipt_CLUSTERIP.c
@@ -29,6 +29,7 @@
 #include <net/netfilter/nf_conntrack.h>
 #include <net/net_namespace.h>
 #include <net/checksum.h>
+#include <net/ip.h>
 #define CLUSTERIP_VERSION "0.8"
@@ -231,24 +232,22 @@ clusterip_hashfn(const struct sk_buff *skb,
 {
        const struct iphdr *iph = ip_hdr(skb);
        unsigned long hashval;
-        u_int16_t sport, dport;
+        u_int16_t sport = 0, dport = 0;
-        const u_int16_t *ports;
+        int poff;
-        switch (iph->protocol) {
+        poff = proto_ports_offset(iph->protocol);
-        case IPPROTO_TCP:
+        if (poff >= 0) {
-        case IPPROTO_UDP:
+                const u_int16_t *ports;
-        case IPPROTO_UDPLITE:
+                u16 _ports[2];
-        case IPPROTO_SCTP:
-        case IPPROTO_DCCP:
+                ports = skb_header_pointer(skb, iph->ihl * 4 + poff, 4, _ports);
-        case IPPROTO_ICMP:
+                if (ports) {
-                ports = (const void *)iph+iph->ihl*4;
+                        sport = ports[0];
-                sport = ports[0];
+                        dport = ports[1];
-                dport = ports[1];
+                }
-                break;
+        } else {
-        default:
                if (net_ratelimit())
                        pr_info("unknown protocol %u\n", iph->protocol);
-                sport = dport = 0;
        }
        switch (config->hash_mode) {
diff --git a/net/ipv4/protocol.c b/net/ipv4/protocol.c
index f2d297351405..65699c24411c 100644
--- a/net/ipv4/protocol.c
+++ b/net/ipv4/protocol.c
@@ -28,8 +28,7 @@
 #include <linux/spinlock.h>
 #include <net/protocol.h>
-const struct net_protocol *inet_protos[MAX_INET_PROTOS] ____cacheline_aligned_in_smp;
+const struct net_protocol *inet_protos[MAX_INET_PROTOS] __read_mostly;
-static DEFINE_SPINLOCK(inet_proto_lock);
 /*
 *      Add a protocol handler to the hash tables
@@ -37,20 +36,9 @@ static DEFINE_SPINLOCK(inet_proto_lock);
 int inet_add_protocol(const struct net_protocol *prot, unsigned char protocol)
 {
-        int hash, ret;
+        int hash = protocol & (MAX_INET_PROTOS - 1);
-        hash = protocol & (MAX_INET_PROTOS - 1);
+        return !cmpxchg(&inet_protos[hash], NULL, prot) ? 0 : -1;
-        spin_lock_bh(&inet_proto_lock);
-        if (inet_protos[hash]) {
-                ret = -1;
-        } else {
-                inet_protos[hash] = prot;
-                ret = 0;
-        }
-        spin_unlock_bh(&inet_proto_lock);
-        return ret;
 }
 EXPORT_SYMBOL(inet_add_protocol);
@@ -60,18 +48,9 @@ EXPORT_SYMBOL(inet_add_protocol);
 int inet_del_protocol(const struct net_protocol *prot, unsigned char protocol)
 {
-        int hash, ret;
+        int ret, hash = protocol & (MAX_INET_PROTOS - 1);
-        hash = protocol & (MAX_INET_PROTOS - 1);
-        spin_lock_bh(&inet_proto_lock);
+        ret = (cmpxchg(&inet_protos[hash], prot, NULL) == prot) ? 0 : -1;
-        if (inet_protos[hash] == prot) {
-                inet_protos[hash] = NULL;
-                ret = 0;
-        } else {
-                ret = -1;
-        }
-        spin_unlock_bh(&inet_proto_lock);
        synchronize_net();
diff --git a/net/ipv4/raw.c b/net/ipv4/raw.c
index 009a7b2aa1ef..1f85ef289895 100644
--- a/net/ipv4/raw.c
+++ b/net/ipv4/raw.c
@@ -505,7 +505,7 @@ static int raw_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
        ipc.addr = inet->inet_saddr;
        ipc.opt = NULL;
-        ipc.shtx.flags = 0;
+        ipc.tx_flags = 0;
        ipc.oif = sk->sk_bound_dev_if;
        if (msg->msg_controllen) {
diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index ac6559cb54f9..04e0df82b88c 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -1107,6 +1107,7 @@ restart:
                 * on the route gc list.
                 */
+                rt->dst.flags |= DST_NOCACHE;
                if (rt->rt_type == RTN_UNICAST || rt->fl.iif == 0) {
                        int err = arp_bind_neighbour(&rt->dst);
                        if (err) {
@@ -1268,18 +1269,11 @@ skip_hashing:
 void rt_bind_peer(struct rtable *rt, int create)
 {
-        static DEFINE_SPINLOCK(rt_peer_lock);
        struct inet_peer *peer;
        peer = inet_getpeer(rt->rt_dst, create);
-        spin_lock_bh(&rt_peer_lock);
+        if (peer && cmpxchg(&rt->peer, NULL, peer) != NULL)
-        if (rt->peer == NULL) {
-                rt->peer = peer;
-                peer = NULL;
-        }
-        spin_unlock_bh(&rt_peer_lock);
-        if (peer)
                inet_putpeer(peer);
 }
@@ -2365,9 +2359,8 @@ static int __mkroute_output(struct rtable **result,
        struct rtable *rth;
        struct in_device *in_dev;
        u32 tos = RT_FL_TOS(oldflp);
-        int err = 0;
-        if (ipv4_is_loopback(fl->fl4_src) && !(dev_out->flags&IFF_LOOPBACK))
+        if (ipv4_is_loopback(fl->fl4_src) && !(dev_out->flags & IFF_LOOPBACK))
                return -EINVAL;
        if (fl->fl4_dst == htonl(0xFFFFFFFF))
@@ -2380,11 +2373,12 @@ static int __mkroute_output(struct rtable **result,
        if (dev_out->flags & IFF_LOOPBACK)
                flags |= RTCF_LOCAL;
-        /* get work reference to inet device */
+        rcu_read_lock();
-        in_dev = in_dev_get(dev_out);
+        in_dev = __in_dev_get_rcu(dev_out);
-        if (!in_dev)
+        if (!in_dev) {
+                rcu_read_unlock();
                return -EINVAL;
+        }
        if (res->type == RTN_BROADCAST) {
                flags |= RTCF_BROADCAST | RTCF_LOCAL;
                if (res->fi) {
@@ -2392,13 +2386,13 @@ static int __mkroute_output(struct rtable **result,
                        res->fi = NULL;
                }
        } else if (res->type == RTN_MULTICAST) {
-                flags |= RTCF_MULTICAST|RTCF_LOCAL;
+                flags |= RTCF_MULTICAST | RTCF_LOCAL;
                if (!ip_check_mc(in_dev, oldflp->fl4_dst, oldflp->fl4_src,
                                 oldflp->proto))
                        flags &= ~RTCF_LOCAL;
                /* If multicast route do not exist use
-                   default one, but do not gateway in this case.
+                 * default one, but do not gateway in this case.
-                   Yes, it is hack.
+                 * Yes, it is hack.
                 */
                if (res->fi && res->prefixlen < 4) {
                        fib_info_put(res->fi);
@@ -2409,9 +2403,12 @@ static int __mkroute_output(struct rtable **result,
        rth = dst_alloc(&ipv4_dst_ops);
        if (!rth) {
-                err = -ENOBUFS;
+                rcu_read_unlock();
-                goto cleanup;
+                return -ENOBUFS;
        }
+        in_dev_hold(in_dev);
+        rcu_read_unlock();
+        rth->idev = in_dev;
        atomic_set(&rth->dst.__refcnt, 1);
        rth->dst.flags= DST_HOST;
@@ -2432,7 +2429,6 @@ static int __mkroute_output(struct rtable **result,
           cache entry */
        rth->dst.dev    = dev_out;
        dev_hold(dev_out);
-        rth->idev       = in_dev_get(dev_out);
        rth->rt_gateway = fl->fl4_dst;
        rth->rt_spec_dst= fl->fl4_src;
@@ -2467,13 +2463,8 @@ static int __mkroute_output(struct rtable **result,
        rt_set_nexthop(rth, res, 0);
        rth->rt_flags = flags;
        *result = rth;
- cleanup:
+        return 0;
-        /* release work reference to inet device */
-        in_dev_put(in_dev);
-        return err;
 }
 static int ip_mkroute_output(struct rtable **rp,
@@ -2497,6 +2488,7 @@ static int ip_mkroute_output(struct rtable **rp,
 /*
 * Major route resolver routine.
+ * called with rcu_read_lock();
 */
 static int ip_route_output_slow(struct net *net, struct rtable **rp,
@@ -2515,7 +2507,7 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
                            .iif = net->loopback_dev->ifindex,
                            .oif = oldflp->oif };
        struct fib_result res;
-        unsigned flags = 0;
+        unsigned int flags = 0;
        struct net_device *dev_out = NULL;
        int free_res = 0;
        int err;
@@ -2545,7 +2537,7 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
                    (ipv4_is_multicast(oldflp->fl4_dst) ||
                     oldflp->fl4_dst == htonl(0xFFFFFFFF))) {
                        /* It is equivalent to inet_addr_type(saddr) == RTN_LOCAL */
-                        dev_out = ip_dev_find(net, oldflp->fl4_src);
+                        dev_out = __ip_dev_find(net, oldflp->fl4_src, false);
                        if (dev_out == NULL)
                                goto out;
@@ -2570,26 +2562,21 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
                if (!(oldflp->flags & FLOWI_FLAG_ANYSRC)) {
                        /* It is equivalent to inet_addr_type(saddr) == RTN_LOCAL */
-                        dev_out = ip_dev_find(net, oldflp->fl4_src);
+                        if (!__ip_dev_find(net, oldflp->fl4_src, false))
-                        if (dev_out == NULL)
                                goto out;
-                        dev_put(dev_out);
-                        dev_out = NULL;
                }
        }
        if (oldflp->oif) {
-                dev_out = dev_get_by_index(net, oldflp->oif);
+                dev_out = dev_get_by_index_rcu(net, oldflp->oif);
                err = -ENODEV;
                if (dev_out == NULL)
                        goto out;
                /* RACE: Check return value of inet_select_addr instead. */
-                if (__in_dev_get_rtnl(dev_out) == NULL) {
+                if (rcu_dereference(dev_out->ip_ptr) == NULL)
-                        dev_put(dev_out);
                        goto out;       /* Wrong error code */
-                }
                if (ipv4_is_local_multicast(oldflp->fl4_dst) ||
                    oldflp->fl4_dst == htonl(0xFFFFFFFF)) {
@@ -2612,10 +2599,7 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
                fl.fl4_dst = fl.fl4_src;
                if (!fl.fl4_dst)
                        fl.fl4_dst = fl.fl4_src = htonl(INADDR_LOOPBACK);
-                if (dev_out)
-                        dev_put(dev_out);
                dev_out = net->loopback_dev;
-                dev_hold(dev_out);
                fl.oif = net->loopback_dev->ifindex;
                res.type = RTN_LOCAL;
                flags |= RTCF_LOCAL;
@@ -2649,8 +2633,6 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
                        res.type = RTN_UNICAST;
                        goto make_route;
                }
-                if (dev_out)
-                        dev_put(dev_out);
                err = -ENETUNREACH;
                goto out;
        }
@@ -2659,10 +2641,7 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
        if (res.type == RTN_LOCAL) {
                if (!fl.fl4_src)
                        fl.fl4_src = fl.fl4_dst;
-                if (dev_out)
-                        dev_put(dev_out);
                dev_out = net->loopback_dev;
-                dev_hold(dev_out);
                fl.oif = dev_out->ifindex;
                if (res.fi)
                        fib_info_put(res.fi);
@@ -2682,28 +2661,23 @@ static int ip_route_output_slow(struct net *net, struct rtable **rp,
        if (!fl.fl4_src)
                fl.fl4_src = FIB_RES_PREFSRC(res);
-        if (dev_out)
-                dev_put(dev_out);
        dev_out = FIB_RES_DEV(res);
-        dev_hold(dev_out);
        fl.oif = dev_out->ifindex;
 make_route:
        err = ip_mkroute_output(rp, &res, &fl, oldflp, dev_out, flags);
        if (free_res)
                fib_res_put(&res);
-        if (dev_out)
-                dev_put(dev_out);
 out:    return err;
 }
 int __ip_route_output_key(struct net *net, struct rtable **rp,
                          const struct flowi *flp)
 {
-        unsigned hash;
+        unsigned int hash;
+        int res;
        struct rtable *rth;
        if (!rt_caching(net))
@@ -2734,7 +2708,10 @@ int __ip_route_output_key(struct net *net, struct rtable **rp,
        rcu_read_unlock_bh();
 slow_output:
-        return ip_route_output_slow(net, rp, flp);
+        rcu_read_lock();
+        res = ip_route_output_slow(net, rp, flp);
+        rcu_read_unlock();
+        return res;
 }
 EXPORT_SYMBOL_GPL(__ip_route_output_key);
@@ -2798,7 +2775,7 @@ static int ipv4_dst_blackhole(struct net *net, struct rtable **rp, struct flowi
        dst_release(&(*rp)->dst);
        *rp = rt;
-        return (rt ? 0 : -ENOMEM);
+        return rt ? 0 : -ENOMEM;
 }
 int ip_route_output_flow(struct net *net, struct rtable **rp, struct flowi *flp,
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
index f115ea68a4ef..1664a0590bb8 100644
--- a/net/ipv4/tcp.c
+++ b/net/ipv4/tcp.c
@@ -2392,7 +2392,12 @@ static int do_tcp_setsockopt(struct sock *sk, int level,
                err = tp->af_specific->md5_parse(sk, optval, optlen);
                break;
 #endif
+        case TCP_USER_TIMEOUT:
+                /* Cap the max timeout in ms TCP will retry/retrans
+                 * before giving up and aborting (ETIMEDOUT) a connection.
+                 */
+                icsk->icsk_user_timeout = msecs_to_jiffies(val);
+                break;
        default:
                err = -ENOPROTOOPT;
                break;
@@ -2611,6 +2616,10 @@ static int do_tcp_getsockopt(struct sock *sk, int level,
        case TCP_THIN_DUPACK:
                val = tp->thin_dupack;
                break;
+        case TCP_USER_TIMEOUT:
+                val = jiffies_to_msecs(icsk->icsk_user_timeout);
+                break;
        default:
                return -ENOPROTOOPT;
        }
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b55f60f6fcbe..f6fdd727a23d 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -182,7 +182,7 @@ static void tcp_incr_quickack(struct sock *sk)
                icsk->icsk_ack.quick = min(quickacks, TCP_MAX_QUICKACKS);
 }
-void tcp_enter_quickack_mode(struct sock *sk)
+static void tcp_enter_quickack_mode(struct sock *sk)
 {
        struct inet_connection_sock *icsk = inet_csk(sk);
        tcp_incr_quickack(sk);
@@ -805,25 +805,12 @@ void tcp_update_metrics(struct sock *sk)
        }
 }
-/* Numbers are taken from RFC3390.
- *
- * John Heffner states:
- *
- *      The RFC specifies a window of no more than 4380 bytes
- *      unless 2*MSS > 4380.  Reading the pseudocode in the RFC
- *      is a bit misleading because they use a clamp at 4380 bytes
- *      rather than use a multiplier in the relevant range.
- */
 __u32 tcp_init_cwnd(struct tcp_sock *tp, struct dst_entry *dst)
 {
        __u32 cwnd = (dst ? dst_metric(dst, RTAX_INITCWND) : 0);
-        if (!cwnd) {
+        if (!cwnd)
-                if (tp->mss_cache > 1460)
+                cwnd = rfc3390_bytes_to_packets(tp->mss_cache);
-                        cwnd = 2;
-                else
-                        cwnd = (tp->mss_cache > 1095) ? 3 : 4;
-        }
        return min_t(__u32, cwnd, tp->snd_cwnd_clamp);
 }
@@ -2314,7 +2301,7 @@ static inline int tcp_dupack_heuristics(struct tcp_sock *tp)
 static inline int tcp_skb_timedout(struct sock *sk, struct sk_buff *skb)
 {
-        return (tcp_time_stamp - TCP_SKB_CB(skb)->when > inet_csk(sk)->icsk_rto);
+        return tcp_time_stamp - TCP_SKB_CB(skb)->when > inet_csk(sk)->icsk_rto;
 }
 static inline int tcp_head_timedout(struct sock *sk)
@@ -3412,8 +3399,8 @@ static void tcp_ack_probe(struct sock *sk)
 static inline int tcp_ack_is_dubious(const struct sock *sk, const int flag)
 {
-        return (!(flag & FLAG_NOT_DUP) || (flag & FLAG_CA_ALERT) ||
+        return !(flag & FLAG_NOT_DUP) || (flag & FLAG_CA_ALERT) ||
-                inet_csk(sk)->icsk_ca_state != TCP_CA_Open);
+                inet_csk(sk)->icsk_ca_state != TCP_CA_Open;
 }
 static inline int tcp_may_raise_cwnd(const struct sock *sk, const int flag)
@@ -3430,9 +3417,9 @@ static inline int tcp_may_update_window(const struct tcp_sock *tp,
                                        const u32 ack, const u32 ack_seq,
                                        const u32 nwin)
 {
-        return (after(ack, tp->snd_una) ||
+        return  after(ack, tp->snd_una) ||
                after(ack_seq, tp->snd_wl1) ||
-                (ack_seq == tp->snd_wl1 && nwin > tp->snd_wnd));
+                (ack_seq == tp->snd_wl1 && nwin > tp->snd_wnd);
 }
 /* Update our send window.
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 020766292bb0..a0232f3a358b 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2571,7 +2571,6 @@ struct sk_buff **tcp4_gro_receive(struct sk_buff **head, struct sk_buff *skb)
        return tcp_gro_receive(head, skb);
 }
-EXPORT_SYMBOL(tcp4_gro_receive);
 int tcp4_gro_complete(struct sk_buff *skb)
 {
@@ -2584,7 +2583,6 @@ int tcp4_gro_complete(struct sk_buff *skb)
        return tcp_gro_complete(skb);
 }
-EXPORT_SYMBOL(tcp4_gro_complete);
 struct proto tcp_prot = {
        .name                   = "TCP",
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index f25b56cb85cb..43cf901d7659 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -55,7 +55,7 @@ static __inline__ int tcp_in_window(u32 seq, u32 end_seq, u32 s_win, u32 e_win)
                return 1;
        if (after(end_seq, s_win) && before(seq, e_win))
                return 1;
-        return (seq == e_win && seq == end_seq);
+        return seq == e_win && seq == end_seq;
 }
 /*
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index de3bd8458588..05b1ecf36763 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -224,16 +224,10 @@ void tcp_select_initial_window(int __space, __u32 mss,
                }
        }
-        /* Set initial window to value enough for senders,
+        /* Set initial window to value enough for senders, following RFC5681. */
-         * following RFC2414. Senders, not following this RFC,
-         * will be satisfied with 2.
-         */
        if (mss > (1 << *rcv_wscale)) {
-                int init_cwnd = 4;
+                int init_cwnd = rfc3390_bytes_to_packets(mss);
-                if (mss > 1460 * 3)
-                        init_cwnd = 2;
-                else if (mss > 1460)
-                        init_cwnd = 3;
                /* when initializing use the value from init_rcv_wnd
                 * rather than the default from above
                 */
@@ -1376,9 +1370,9 @@ static inline int tcp_nagle_check(const struct tcp_sock *tp,
                                  const struct sk_buff *skb,
                                  unsigned mss_now, int nonagle)
 {
-        return (skb->len < mss_now &&
+        return skb->len < mss_now &&
                ((nonagle & TCP_NAGLE_CORK) ||
-                 (!nonagle && tp->packets_out && tcp_minshall_check(tp))));
+                 (!nonagle && tp->packets_out && tcp_minshall_check(tp)));
 }
 /* Return non-zero if the Nagle test allows this packet to be
@@ -1449,10 +1443,10 @@ int tcp_may_send_now(struct sock *sk)
        struct tcp_sock *tp = tcp_sk(sk);
        struct sk_buff *skb = tcp_send_head(sk);
-        return (skb &&
+        return skb &&
                tcp_snd_test(sk, skb, tcp_current_mss(sk),
                             (tcp_skb_is_last(sk, skb) ?
-                              tp->nonagle : TCP_NAGLE_PUSH)));
+                              tp->nonagle : TCP_NAGLE_PUSH));
 }
 /* Trim TSO SKB to LEN bytes, put the remaining data into a new packet
@@ -2429,6 +2423,12 @@ struct sk_buff *tcp_make_synack(struct sock *sk, struct dst_entry *dst,
                __u8 rcv_wscale;
                /* Set this up on the first call only */
                req->window_clamp = tp->window_clamp ? : dst_metric(dst, RTAX_WINDOW);
+                /* limit the window selection if the user enforce a smaller rx buffer */
+                if (sk->sk_userlocks & SOCK_RCVBUF_LOCK &&
+                    (req->window_clamp > tcp_full_space(sk) || req->window_clamp == 0))
+                        req->window_clamp = tcp_full_space(sk);
                /* tcp_full_space because it is guaranteed to be the first packet */
                tcp_select_initial_window(tcp_full_space(sk),
                        mss - (ireq->tstamp_ok ? TCPOLEN_TSTAMP_ALIGNED : 0),
@@ -2555,6 +2555,11 @@ static void tcp_connect_init(struct sock *sk)
        tcp_initialize_rcv_mss(sk);
+        /* limit the window selection if the user enforce a smaller rx buffer */
+        if (sk->sk_userlocks & SOCK_RCVBUF_LOCK &&
+            (tp->window_clamp > tcp_full_space(sk) || tp->window_clamp == 0))
+                tp->window_clamp = tcp_full_space(sk);
        tcp_select_initial_window(tcp_full_space(sk),
                                  tp->advmss - (tp->rx_opt.ts_recent_stamp ? tp->tcp_header_len - sizeof(struct tcphdr) : 0),
                                  &tp->rcv_wnd,
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 74c54b30600f..f3c8c6c019ae 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -140,10 +140,10 @@ static void tcp_mtu_probing(struct inet_connection_sock *icsk, struct sock *sk)
 */
 static bool retransmits_timed_out(struct sock *sk,
                                  unsigned int boundary,
+                                  unsigned int timeout,
                                  bool syn_set)
 {
-        unsigned int timeout, linear_backoff_thresh;
+        unsigned int linear_backoff_thresh, start_ts;
-        unsigned int start_ts;
        unsigned int rto_base = syn_set ? TCP_TIMEOUT_INIT : TCP_RTO_MIN;
        if (!inet_csk(sk)->icsk_retransmits)
@@ -154,14 +154,15 @@ static bool retransmits_timed_out(struct sock *sk,
        else
                start_ts = tcp_sk(sk)->retrans_stamp;
-        linear_backoff_thresh = ilog2(TCP_RTO_MAX/rto_base);
+        if (likely(timeout == 0)) {
+                linear_backoff_thresh = ilog2(TCP_RTO_MAX/rto_base);
-        if (boundary <= linear_backoff_thresh)
-                timeout = ((2 << boundary) - 1) * rto_base;
-        else
-                timeout = ((2 << linear_backoff_thresh) - 1) * rto_base +
-                          (boundary - linear_backoff_thresh) * TCP_RTO_MAX;
+                if (boundary <= linear_backoff_thresh)
+                        timeout = ((2 << boundary) - 1) * rto_base;
+                else
+                        timeout = ((2 << linear_backoff_thresh) - 1) * rto_base +
+                                (boundary - linear_backoff_thresh) * TCP_RTO_MAX;
+        }
        return (tcp_time_stamp - start_ts) >= timeout;
 }
@@ -178,7 +179,7 @@ static int tcp_write_timeout(struct sock *sk)
                retry_until = icsk->icsk_syn_retries ? : sysctl_tcp_syn_retries;
                syn_set = 1;
        } else {
-                if (retransmits_timed_out(sk, sysctl_tcp_retries1, 0)) {
+                if (retransmits_timed_out(sk, sysctl_tcp_retries1, 0, 0)) {
                        /* Black hole detection */
                        tcp_mtu_probing(icsk, sk);
@@ -191,14 +192,15 @@ static int tcp_write_timeout(struct sock *sk)
                        retry_until = tcp_orphan_retries(sk, alive);
                        do_reset = alive ||
-                                   !retransmits_timed_out(sk, retry_until, 0);
+                                !retransmits_timed_out(sk, retry_until, 0, 0);
                        if (tcp_out_of_resources(sk, do_reset))
                                return 1;
                }
        }
-        if (retransmits_timed_out(sk, retry_until, syn_set)) {
+        if (retransmits_timed_out(sk, retry_until,
+                                  syn_set ? 0 : icsk->icsk_user_timeout, syn_set)) {
                /* Has it gone just too far? */
                tcp_write_err(sk);
                return 1;
@@ -440,7 +442,7 @@ out_reset_timer:
                icsk->icsk_rto = min(icsk->icsk_rto << 1, TCP_RTO_MAX);
        }
        inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS, icsk->icsk_rto, TCP_RTO_MAX);
-        if (retransmits_timed_out(sk, sysctl_tcp_retries1 + 1, 0))
+        if (retransmits_timed_out(sk, sysctl_tcp_retries1 + 1, 0, 0))
                __sk_dst_reset(sk);
 out:;
@@ -560,7 +562,14 @@ static void tcp_keepalive_timer (unsigned long data)
        elapsed = keepalive_time_elapsed(tp);
        if (elapsed >= keepalive_time_when(tp)) {
-                if (icsk->icsk_probes_out >= keepalive_probes(tp)) {
+                /* If the TCP_USER_TIMEOUT option is enabled, use that
+                 * to determine when to timeout instead.
+                 */
+                if ((icsk->icsk_user_timeout != 0 &&
+                    elapsed >= icsk->icsk_user_timeout &&
+                    icsk->icsk_probes_out > 0) ||
+                    (icsk->icsk_user_timeout == 0 &&
+                    icsk->icsk_probes_out >= keepalive_probes(tp))) {
                        tcp_send_active_reset(sk, GFP_ATOMIC);
                        tcp_write_err(sk);
                        goto out;
diff --git a/net/ipv4/tcp_westwood.c b/net/ipv4/tcp_westwood.c
index 20151d6a6241..a534dda5456e 100644
--- a/net/ipv4/tcp_westwood.c
+++ b/net/ipv4/tcp_westwood.c
@@ -80,7 +80,7 @@ static void tcp_westwood_init(struct sock *sk)
 */
 static inline u32 westwood_do_filter(u32 a, u32 b)
 {
-        return (((7 * a) + b) >> 3);
+        return ((7 * a) + b) >> 3;
 }
 static void westwood_filter(struct westwood *w, u32 delta)
diff --git a/net/ipv4/tunnel4.c b/net/ipv4/tunnel4.c
index 59186ca7808a..9a17bd2a0a37 100644
--- a/net/ipv4/tunnel4.c
+++ b/net/ipv4/tunnel4.c
@@ -14,8 +14,8 @@
 #include <net/protocol.h>
 #include <net/xfrm.h>
-static struct xfrm_tunnel *tunnel4_handlers;
+static struct xfrm_tunnel *tunnel4_handlers __read_mostly;
-static struct xfrm_tunnel *tunnel64_handlers;
+static struct xfrm_tunnel *tunnel64_handlers __read_mostly;
 static DEFINE_MUTEX(tunnel4_mutex);
 static inline struct xfrm_tunnel **fam_handlers(unsigned short family)
@@ -39,7 +39,7 @@ int xfrm4_tunnel_register(struct xfrm_tunnel *handler, unsigned short family)
        }
        handler->next = *pprev;
-        *pprev = handler;
+        rcu_assign_pointer(*pprev, handler);
        ret = 0;
@@ -73,6 +73,11 @@ int xfrm4_tunnel_deregister(struct xfrm_tunnel *handler, unsigned short family)
 }
 EXPORT_SYMBOL(xfrm4_tunnel_deregister);
+#define for_each_tunnel_rcu(head, handler)              \
+        for (handler = rcu_dereference(head);           \
+             handler != NULL;                           \
+             handler = rcu_dereference(handler->next))  \
+        
 static int tunnel4_rcv(struct sk_buff *skb)
 {
        struct xfrm_tunnel *handler;
@@ -80,7 +85,7 @@ static int tunnel4_rcv(struct sk_buff *skb)
        if (!pskb_may_pull(skb, sizeof(struct iphdr)))
                goto drop;
-        for (handler = tunnel4_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel4_handlers, handler)
                if (!handler->handler(skb))
                        return 0;
@@ -99,7 +104,7 @@ static int tunnel64_rcv(struct sk_buff *skb)
        if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
                goto drop;
-        for (handler = tunnel64_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel64_handlers, handler)
                if (!handler->handler(skb))
                        return 0;
@@ -115,7 +120,7 @@ static void tunnel4_err(struct sk_buff *skb, u32 info)
 {
        struct xfrm_tunnel *handler;
-        for (handler = tunnel4_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel4_handlers, handler)
                if (!handler->err_handler(skb, info))
                        break;
 }
@@ -125,7 +130,7 @@ static void tunnel64_err(struct sk_buff *skb, u32 info)
 {
        struct xfrm_tunnel *handler;
-        for (handler = tunnel64_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel64_handlers, handler)
                if (!handler->err_handler(skb, info))
                        break;
 }
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index fb23c2e63b52..b3f7e8cf18ac 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -797,7 +797,7 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
                return -EOPNOTSUPP;
        ipc.opt = NULL;
-        ipc.shtx.flags = 0;
+        ipc.tx_flags = 0;
        if (up->pending) {
                /*
@@ -845,7 +845,7 @@ int udp_sendmsg(struct kiocb *iocb, struct sock *sk, struct msghdr *msg,
        ipc.addr = inet->inet_saddr;
        ipc.oif = sk->sk_bound_dev_if;
-        err = sock_tx_timestamp(msg, sk, &ipc.shtx);
+        err = sock_tx_timestamp(sk, &ipc.tx_flags);
        if (err)
                return err;
        if (msg->msg_controllen) {
diff --git a/net/ipv4/xfrm4_tunnel.c b/net/ipv4/xfrm4_tunnel.c
index 41f5982d2087..82806455e859 100644
--- a/net/ipv4/xfrm4_tunnel.c
+++ b/net/ipv4/xfrm4_tunnel.c
@@ -58,14 +58,14 @@ static int xfrm_tunnel_err(struct sk_buff *skb, u32 info)
        return -ENOENT;
 }
-static struct xfrm_tunnel xfrm_tunnel_handler = {
+static struct xfrm_tunnel xfrm_tunnel_handler __read_mostly = {
        .handler        =       xfrm_tunnel_rcv,
        .err_handler    =       xfrm_tunnel_err,
        .priority       =       2,
 };
 #if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
-static struct xfrm_tunnel xfrm64_tunnel_handler = {
+static struct xfrm_tunnel xfrm64_tunnel_handler __read_mostly = {
        .handler        =       xfrm_tunnel_rcv,
        .err_handler    =       xfrm_tunnel_err,
        .priority       =       2,
diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 324fac3b6c16..8c88340278f5 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -243,7 +243,7 @@ static inline bool addrconf_qdisc_ok(const struct net_device *dev)
 /* Check if a route is valid prefix route */
 static inline int addrconf_is_prefix_route(const struct rt6_info *rt)
 {
-        return ((rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0);
+        return (rt->rt6i_flags & (RTF_GATEWAY | RTF_DEFAULT)) == 0;
 }
 static void addrconf_del_timer(struct inet6_ifaddr *ifp)
@@ -2964,7 +2964,8 @@ static void addrconf_dad_completed(struct inet6_ifaddr *ifp)
           start sending router solicitations.
         */
-        if (ifp->idev->cnf.forwarding == 0 &&
+        if ((ifp->idev->cnf.forwarding == 0 ||
+             ifp->idev->cnf.forwarding == 2) &&
            ifp->idev->cnf.rtr_solicits > 0 &&
            (dev->flags&IFF_LOOPBACK) == 0 &&
            (ipv6_addr_type(&ifp->addr) & IPV6_ADDR_LINKLOCAL)) {
diff --git a/net/ipv6/addrlabel.c b/net/ipv6/addrlabel.c
index 8175f802651b..c8993e5a337c 100644
--- a/net/ipv6/addrlabel.c
+++ b/net/ipv6/addrlabel.c
@@ -518,10 +518,9 @@ static int ip6addrlbl_dump(struct sk_buff *skb, struct netlink_callback *cb)
 static inline int ip6addrlbl_msgsize(void)
 {
-        return (NLMSG_ALIGN(sizeof(struct ifaddrlblmsg))
+        return NLMSG_ALIGN(sizeof(struct ifaddrlblmsg))
                + nla_total_size(16)    /* IFAL_ADDRESS */
-                + nla_total_size(4)     /* IFAL_LABEL */
+                + nla_total_size(4);    /* IFAL_LABEL */
-        );
 }
 static int ip6addrlbl_get(struct sk_buff *in_skb, struct nlmsghdr* nlh,
diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index 56b9bf2516f4..60220985bb80 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -467,7 +467,7 @@ int inet6_getname(struct socket *sock, struct sockaddr *uaddr,
        if (ipv6_addr_type(&sin->sin6_addr) & IPV6_ADDR_LINKLOCAL)
                sin->sin6_scope_id = sk->sk_bound_dev_if;
        *uaddr_len = sizeof(*sin);
-        return(0);
+        return 0;
 }
 EXPORT_SYMBOL(inet6_getname);
@@ -488,7 +488,7 @@ int inet6_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
        case SIOCADDRT:
        case SIOCDELRT:
-                return(ipv6_route_ioctl(net, cmd, (void __user *)arg));
+                return ipv6_route_ioctl(net, cmd, (void __user *)arg);
        case SIOCSIFADDR:
                return addrconf_add_ifaddr(net, (void __user *) arg);
@@ -502,7 +502,7 @@ int inet6_ioctl(struct socket *sock, unsigned int cmd, unsigned long arg)
                return sk->sk_prot->ioctl(sk, cmd, arg);
        }
        /*NOTREACHED*/
-        return(0);
+        return 0;
 }
 EXPORT_SYMBOL(inet6_ioctl);
diff --git a/net/ipv6/exthdrs_core.c b/net/ipv6/exthdrs_core.c
index e1caa5d526c2..14ed0a955b56 100644
--- a/net/ipv6/exthdrs_core.c
+++ b/net/ipv6/exthdrs_core.c
@@ -13,12 +13,12 @@ int ipv6_ext_hdr(u8 nexthdr)
        /*
         * find out if nexthdr is an extension header or a protocol
         */
-        return ( (nexthdr == NEXTHDR_HOP)       ||
+        return   (nexthdr == NEXTHDR_HOP)       ||
                 (nexthdr == NEXTHDR_ROUTING)   ||
                 (nexthdr == NEXTHDR_FRAGMENT)  ||
                 (nexthdr == NEXTHDR_AUTH)      ||
                 (nexthdr == NEXTHDR_NONE)      ||
-                 (nexthdr == NEXTHDR_DEST) );
+                 (nexthdr == NEXTHDR_DEST);
 }
 /*
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 980912ed7a38..99157b4cd56e 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -637,7 +637,7 @@ static int ip6_fragment(struct sk_buff *skb, int (*output)(struct sk_buff *))
        }
        mtu -= hlen + sizeof(struct frag_hdr);
-        if (skb_has_frags(skb)) {
+        if (skb_has_frag_list(skb)) {
                int first_len = skb_pagelen(skb);
                struct sk_buff *frag2;
@@ -878,8 +878,8 @@ static inline int ip6_rt_check(struct rt6key *rt_key,
                               struct in6_addr *fl_addr,
                               struct in6_addr *addr_cache)
 {
-        return ((rt_key->plen != 128 || !ipv6_addr_equal(fl_addr, &rt_key->addr)) &&
+        return (rt_key->plen != 128 || !ipv6_addr_equal(fl_addr, &rt_key->addr)) &&
-                (addr_cache == NULL || !ipv6_addr_equal(fl_addr, addr_cache)));
+                (addr_cache == NULL || !ipv6_addr_equal(fl_addr, addr_cache));
 }
 static struct dst_entry *ip6_sk_dst_check(struct sock *sk,
diff --git a/net/ipv6/ip6_tunnel.c b/net/ipv6/ip6_tunnel.c
index 0fd027f3f47e..8be3c452af90 100644
--- a/net/ipv6/ip6_tunnel.c
+++ b/net/ipv6/ip6_tunnel.c
@@ -75,7 +75,7 @@ MODULE_LICENSE("GPL");
                     (addr)->s6_addr32[2] ^ (addr)->s6_addr32[3]) & \
                    (HASH_SIZE - 1))
-static void ip6_tnl_dev_init(struct net_device *dev);
+static int ip6_tnl_dev_init(struct net_device *dev);
 static void ip6_tnl_dev_setup(struct net_device *dev);
 static int ip6_tnl_net_id __read_mostly;
@@ -83,15 +83,42 @@ struct ip6_tnl_net {
        /* the IPv6 tunnel fallback device */
        struct net_device *fb_tnl_dev;
        /* lists for storing tunnels in use */
-        struct ip6_tnl *tnls_r_l[HASH_SIZE];
+        struct ip6_tnl __rcu *tnls_r_l[HASH_SIZE];
-        struct ip6_tnl *tnls_wc[1];
+        struct ip6_tnl __rcu *tnls_wc[1];
-        struct ip6_tnl **tnls[2];
+        struct ip6_tnl __rcu **tnls[2];
 };
+/* often modified stats are per cpu, other are shared (netdev->stats) */
+struct pcpu_tstats {
+        unsigned long   rx_packets;
+        unsigned long   rx_bytes;
+        unsigned long   tx_packets;
+        unsigned long   tx_bytes;
+};
+static struct net_device_stats *ip6_get_stats(struct net_device *dev)
+{
+        struct pcpu_tstats sum = { 0 };
+        int i;
+        for_each_possible_cpu(i) {
+                const struct pcpu_tstats *tstats = per_cpu_ptr(dev->tstats, i);
+                sum.rx_packets += tstats->rx_packets;
+                sum.rx_bytes   += tstats->rx_bytes;
+                sum.tx_packets += tstats->tx_packets;
+                sum.tx_bytes   += tstats->tx_bytes;
+        }
+        dev->stats.rx_packets = sum.rx_packets;
+        dev->stats.rx_bytes   = sum.rx_bytes;
+        dev->stats.tx_packets = sum.tx_packets;
+        dev->stats.tx_bytes   = sum.tx_bytes;
+        return &dev->stats;
+}
 /*
- * Locking : hash tables are protected by RCU and a spinlock
+ * Locking : hash tables are protected by RCU and RTNL
 */
-static DEFINE_SPINLOCK(ip6_tnl_lock);
 static inline struct dst_entry *ip6_tnl_dst_check(struct ip6_tnl *t)
 {
@@ -138,8 +165,8 @@ static inline void ip6_tnl_dst_store(struct ip6_tnl *t, struct dst_entry *dst)
 static struct ip6_tnl *
 ip6_tnl_lookup(struct net *net, struct in6_addr *remote, struct in6_addr *local)
 {
-        unsigned h0 = HASH(remote);
+        unsigned int h0 = HASH(remote);
-        unsigned h1 = HASH(local);
+        unsigned int h1 = HASH(local);
        struct ip6_tnl *t;
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
@@ -167,7 +194,7 @@ ip6_tnl_lookup(struct net *net, struct in6_addr *remote, struct in6_addr *local)
 * Return: head of IPv6 tunnel list
 **/
-static struct ip6_tnl **
+static struct ip6_tnl __rcu **
 ip6_tnl_bucket(struct ip6_tnl_net *ip6n, struct ip6_tnl_parm *p)
 {
        struct in6_addr *remote = &p->raddr;
@@ -190,12 +217,10 @@ ip6_tnl_bucket(struct ip6_tnl_net *ip6n, struct ip6_tnl_parm *p)
 static void
 ip6_tnl_link(struct ip6_tnl_net *ip6n, struct ip6_tnl *t)
 {
-        struct ip6_tnl **tp = ip6_tnl_bucket(ip6n, &t->parms);
+        struct ip6_tnl __rcu **tp = ip6_tnl_bucket(ip6n, &t->parms);
-        spin_lock_bh(&ip6_tnl_lock);
+        rcu_assign_pointer(t->next , rtnl_dereference(*tp));
-        t->next = *tp;
        rcu_assign_pointer(*tp, t);
-        spin_unlock_bh(&ip6_tnl_lock);
 }
 /**
@@ -206,18 +231,25 @@ ip6_tnl_link(struct ip6_tnl_net *ip6n, struct ip6_tnl *t)
 static void
 ip6_tnl_unlink(struct ip6_tnl_net *ip6n, struct ip6_tnl *t)
 {
-        struct ip6_tnl **tp;
+        struct ip6_tnl __rcu **tp;
+        struct ip6_tnl *iter;
-        for (tp = ip6_tnl_bucket(ip6n, &t->parms); *tp; tp = &(*tp)->next) {
-                if (t == *tp) {
+        for (tp = ip6_tnl_bucket(ip6n, &t->parms);
-                        spin_lock_bh(&ip6_tnl_lock);
+             (iter = rtnl_dereference(*tp)) != NULL;
-                        *tp = t->next;
+             tp = &iter->next) {
-                        spin_unlock_bh(&ip6_tnl_lock);
+                if (t == iter) {
+                        rcu_assign_pointer(*tp, t->next);
                        break;
                }
        }
 }
+static void ip6_dev_free(struct net_device *dev)
+{
+        free_percpu(dev->tstats);
+        free_netdev(dev);
+}
 /**
 * ip6_tnl_create() - create a new tunnel
 *   @p: tunnel parameters
@@ -256,7 +288,9 @@ static struct ip6_tnl *ip6_tnl_create(struct net *net, struct ip6_tnl_parm *p)
        t = netdev_priv(dev);
        t->parms = *p;
-        ip6_tnl_dev_init(dev);
+        err = ip6_tnl_dev_init(dev);
+        if (err < 0)
+                goto failed_free;
        if ((err = register_netdevice(dev)) < 0)
                goto failed_free;
@@ -266,7 +300,7 @@ static struct ip6_tnl *ip6_tnl_create(struct net *net, struct ip6_tnl_parm *p)
        return t;
 failed_free:
-        free_netdev(dev);
+        ip6_dev_free(dev);
 failed:
        return NULL;
 }
@@ -290,10 +324,13 @@ static struct ip6_tnl *ip6_tnl_locate(struct net *net,
 {
        struct in6_addr *remote = &p->raddr;
        struct in6_addr *local = &p->laddr;
+        struct ip6_tnl __rcu **tp;
        struct ip6_tnl *t;
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
-        for (t = *ip6_tnl_bucket(ip6n, p); t; t = t->next) {
+        for (tp = ip6_tnl_bucket(ip6n, p);
+             (t = rtnl_dereference(*tp)) != NULL;
+             tp = &t->next) {
                if (ipv6_addr_equal(local, &t->parms.laddr) &&
                    ipv6_addr_equal(remote, &t->parms.raddr))
                        return t;
@@ -318,13 +355,10 @@ ip6_tnl_dev_uninit(struct net_device *dev)
        struct net *net = dev_net(dev);
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
-        if (dev == ip6n->fb_tnl_dev) {
+        if (dev == ip6n->fb_tnl_dev)
-                spin_lock_bh(&ip6_tnl_lock);
+                rcu_assign_pointer(ip6n->tnls_wc[0], NULL);
-                ip6n->tnls_wc[0] = NULL;
+        else
-                spin_unlock_bh(&ip6_tnl_lock);
-        } else {
                ip6_tnl_unlink(ip6n, t);
-        }
        ip6_tnl_dst_reset(t);
        dev_put(dev);
 }
@@ -702,6 +736,8 @@ static int ip6_tnl_rcv(struct sk_buff *skb, __u16 protocol,
        if ((t = ip6_tnl_lookup(dev_net(skb->dev), &ipv6h->saddr,
                                        &ipv6h->daddr)) != NULL) {
+                struct pcpu_tstats *tstats;
                if (t->parms.proto != ipproto && t->parms.proto != 0) {
                        rcu_read_unlock();
                        goto discard;
@@ -724,10 +760,17 @@ static int ip6_tnl_rcv(struct sk_buff *skb, __u16 protocol,
                skb->pkt_type = PACKET_HOST;
                memset(skb->cb, 0, sizeof(struct inet6_skb_parm));
-                skb_tunnel_rx(skb, t->dev);
+                tstats = this_cpu_ptr(t->dev->tstats);
+                tstats->rx_packets++;
+                tstats->rx_bytes += skb->len;
+                __skb_tunnel_rx(skb, t->dev);
                dscp_ecn_decapsulate(t, ipv6h, skb);
-                netif_rx(skb);
+                if (netif_rx(skb) == NET_RX_DROP)
+                        t->dev->stats.rx_dropped++;
                rcu_read_unlock();
                return 0;
        }
@@ -934,8 +977,10 @@ static int ip6_tnl_xmit2(struct sk_buff *skb,
        err = ip6_local_out(skb);
        if (net_xmit_eval(err) == 0) {
-                stats->tx_bytes += pkt_len;
+                struct pcpu_tstats *tstats = this_cpu_ptr(t->dev->tstats);
-                stats->tx_packets++;
+                tstats->tx_bytes += pkt_len;
+                tstats->tx_packets++;
        } else {
                stats->tx_errors++;
                stats->tx_aborted_errors++;
@@ -1300,12 +1345,14 @@ ip6_tnl_change_mtu(struct net_device *dev, int new_mtu)
 static const struct net_device_ops ip6_tnl_netdev_ops = {
-        .ndo_uninit = ip6_tnl_dev_uninit,
+        .ndo_uninit     = ip6_tnl_dev_uninit,
        .ndo_start_xmit = ip6_tnl_xmit,
-        .ndo_do_ioctl = ip6_tnl_ioctl,
+        .ndo_do_ioctl   = ip6_tnl_ioctl,
        .ndo_change_mtu = ip6_tnl_change_mtu,
+        .ndo_get_stats  = ip6_get_stats,
 };
 /**
 * ip6_tnl_dev_setup - setup virtual tunnel device
 *   @dev: virtual device associated with tunnel
@@ -1317,7 +1364,7 @@ static const struct net_device_ops ip6_tnl_netdev_ops = {
 static void ip6_tnl_dev_setup(struct net_device *dev)
 {
        dev->netdev_ops = &ip6_tnl_netdev_ops;
-        dev->destructor = free_netdev;
+        dev->destructor = ip6_dev_free;
        dev->type = ARPHRD_TUNNEL6;
        dev->hard_header_len = LL_MAX_HEADER + sizeof (struct ipv6hdr);
@@ -1333,12 +1380,17 @@ static void ip6_tnl_dev_setup(struct net_device *dev)
 *   @dev: virtual device associated with tunnel
 **/
-static inline void
+static inline int
 ip6_tnl_dev_init_gen(struct net_device *dev)
 {
        struct ip6_tnl *t = netdev_priv(dev);
        t->dev = dev;
        strcpy(t->parms.name, dev->name);
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
+        return 0;
 }
 /**
@@ -1346,11 +1398,15 @@ ip6_tnl_dev_init_gen(struct net_device *dev)
 *   @dev: virtual device associated with tunnel
 **/
-static void ip6_tnl_dev_init(struct net_device *dev)
+static int ip6_tnl_dev_init(struct net_device *dev)
 {
        struct ip6_tnl *t = netdev_priv(dev);
-        ip6_tnl_dev_init_gen(dev);
+        int err = ip6_tnl_dev_init_gen(dev);
+        if (err)
+                return err;
        ip6_tnl_link_config(t);
+        return 0;
 }
 /**
@@ -1360,25 +1416,29 @@ static void ip6_tnl_dev_init(struct net_device *dev)
 * Return: 0
 **/
-static void __net_init ip6_fb_tnl_dev_init(struct net_device *dev)
+static int __net_init ip6_fb_tnl_dev_init(struct net_device *dev)
 {
        struct ip6_tnl *t = netdev_priv(dev);
        struct net *net = dev_net(dev);
        struct ip6_tnl_net *ip6n = net_generic(net, ip6_tnl_net_id);
+        int err = ip6_tnl_dev_init_gen(dev);
+        if (err)
+                return err;
-        ip6_tnl_dev_init_gen(dev);
        t->parms.proto = IPPROTO_IPV6;
        dev_hold(dev);
-        ip6n->tnls_wc[0] = t;
+        rcu_assign_pointer(ip6n->tnls_wc[0], t);
+        return 0;
 }
-static struct xfrm6_tunnel ip4ip6_handler = {
+static struct xfrm6_tunnel ip4ip6_handler __read_mostly = {
        .handler        = ip4ip6_rcv,
        .err_handler    = ip4ip6_err,
        .priority       =       1,
 };
-static struct xfrm6_tunnel ip6ip6_handler = {
+static struct xfrm6_tunnel ip6ip6_handler __read_mostly = {
        .handler        = ip6ip6_rcv,
        .err_handler    = ip6ip6_err,
        .priority       =       1,
@@ -1391,14 +1451,14 @@ static void __net_exit ip6_tnl_destroy_tunnels(struct ip6_tnl_net *ip6n)
        LIST_HEAD(list);
        for (h = 0; h < HASH_SIZE; h++) {
-                t = ip6n->tnls_r_l[h];
+                t = rtnl_dereference(ip6n->tnls_r_l[h]);
                while (t != NULL) {
                        unregister_netdevice_queue(t->dev, &list);
-                        t = t->next;
+                        t = rtnl_dereference(t->next);
                }
        }
-        t = ip6n->tnls_wc[0];
+        t = rtnl_dereference(ip6n->tnls_wc[0]);
        unregister_netdevice_queue(t->dev, &list);
        unregister_netdevice_many(&list);
 }
@@ -1419,7 +1479,9 @@ static int __net_init ip6_tnl_init_net(struct net *net)
                goto err_alloc_dev;
        dev_net_set(ip6n->fb_tnl_dev, net);
-        ip6_fb_tnl_dev_init(ip6n->fb_tnl_dev);
+        err = ip6_fb_tnl_dev_init(ip6n->fb_tnl_dev);
+        if (err < 0)
+                goto err_register;
        err = register_netdev(ip6n->fb_tnl_dev);
        if (err < 0)
@@ -1427,7 +1489,7 @@ static int __net_init ip6_tnl_init_net(struct net *net)
        return 0;
 err_register:
-        free_netdev(ip6n->fb_tnl_dev);
+        ip6_dev_free(ip6n->fb_tnl_dev);
 err_alloc_dev:
        return err;
 }
diff --git a/net/ipv6/ip6mr.c b/net/ipv6/ip6mr.c
index 66078dad7fe8..2640c9be589d 100644
--- a/net/ipv6/ip6mr.c
+++ b/net/ipv6/ip6mr.c
@@ -666,7 +666,9 @@ static int pim6_rcv(struct sk_buff *skb)
        skb_tunnel_rx(skb, reg_dev);
-        netif_rx(skb);
+        if (netif_rx(skb) == NET_RX_DROP)
+                reg_dev->stats.rx_dropped++;
        dev_put(reg_dev);
        return 0;
 drop:
diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 58841c4ae947..b3dd844cd34f 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -228,12 +228,12 @@ static struct nd_opt_hdr *ndisc_next_option(struct nd_opt_hdr *cur,
        do {
                cur = ((void *)cur) + (cur->nd_opt_len << 3);
        } while(cur < end && cur->nd_opt_type != type);
-        return (cur <= end && cur->nd_opt_type == type ? cur : NULL);
+        return cur <= end && cur->nd_opt_type == type ? cur : NULL;
 }
 static inline int ndisc_is_useropt(struct nd_opt_hdr *opt)
 {
-        return (opt->nd_opt_type == ND_OPT_RDNSS);
+        return opt->nd_opt_type == ND_OPT_RDNSS;
 }
 static struct nd_opt_hdr *ndisc_next_useropt(struct nd_opt_hdr *cur,
@@ -244,7 +244,7 @@ static struct nd_opt_hdr *ndisc_next_useropt(struct nd_opt_hdr *cur,
        do {
                cur = ((void *)cur) + (cur->nd_opt_len << 3);
        } while(cur < end && !ndisc_is_useropt(cur));
-        return (cur <= end && ndisc_is_useropt(cur) ? cur : NULL);
+        return cur <= end && ndisc_is_useropt(cur) ? cur : NULL;
 }
 static struct ndisc_options *ndisc_parse_options(u8 *opt, int opt_len,
@@ -319,7 +319,7 @@ static inline u8 *ndisc_opt_addr_data(struct nd_opt_hdr *p,
        int prepad = ndisc_addr_option_pad(dev->type);
        if (lladdrlen != NDISC_OPT_SPACE(dev->addr_len + prepad))
                return NULL;
-        return (lladdr + prepad);
+        return lladdr + prepad;
 }
 int ndisc_mc_map(struct in6_addr *addr, char *buf, struct net_device *dev, int dir)
@@ -1105,6 +1105,18 @@ errout:
        rtnl_set_sk_err(net, RTNLGRP_ND_USEROPT, err);
 }
+static inline int accept_ra(struct inet6_dev *in6_dev)
+{
+        /*
+         * If forwarding is enabled, RA are not accepted unless the special
+         * hybrid mode (accept_ra=2) is enabled.
+         */
+        if (in6_dev->cnf.forwarding && in6_dev->cnf.accept_ra < 2)
+                return 0;
+        return in6_dev->cnf.accept_ra;
+}
 static void ndisc_router_discovery(struct sk_buff *skb)
 {
        struct ra_msg *ra_msg = (struct ra_msg *)skb_transport_header(skb);
@@ -1158,8 +1170,7 @@ static void ndisc_router_discovery(struct sk_buff *skb)
                return;
        }
-        /* skip route and link configuration on routers */
+        if (!accept_ra(in6_dev))
-        if (in6_dev->cnf.forwarding || !in6_dev->cnf.accept_ra)
                goto skip_linkparms;
 #ifdef CONFIG_IPV6_NDISC_NODETYPE
@@ -1309,8 +1320,7 @@ skip_linkparms:
                             NEIGH_UPDATE_F_ISROUTER);
        }
-        /* skip route and link configuration on routers */
+        if (!accept_ra(in6_dev))
-        if (in6_dev->cnf.forwarding || !in6_dev->cnf.accept_ra)
                goto out;
 #ifdef CONFIG_IPV6_ROUTE_INFO
diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c
index 8e754be92c24..6b331e9b5706 100644
--- a/net/ipv6/netfilter/ip6_tables.c
+++ b/net/ipv6/netfilter/ip6_tables.c
@@ -82,13 +82,13 @@ EXPORT_SYMBOL_GPL(ip6t_alloc_initial_table);
 int
 ip6t_ext_hdr(u8 nexthdr)
 {
-        return ( (nexthdr == IPPROTO_HOPOPTS)   ||
+        return  (nexthdr == IPPROTO_HOPOPTS)   ||
-                 (nexthdr == IPPROTO_ROUTING)   ||
+                (nexthdr == IPPROTO_ROUTING)   ||
-                 (nexthdr == IPPROTO_FRAGMENT)  ||
+                (nexthdr == IPPROTO_FRAGMENT)  ||
-                 (nexthdr == IPPROTO_ESP)       ||
+                (nexthdr == IPPROTO_ESP)       ||
-                 (nexthdr == IPPROTO_AH)        ||
+                (nexthdr == IPPROTO_AH)        ||
-                 (nexthdr == IPPROTO_NONE)      ||
+                (nexthdr == IPPROTO_NONE)      ||
-                 (nexthdr == IPPROTO_DSTOPTS) );
+                (nexthdr == IPPROTO_DSTOPTS);
 }
 /* Returns whether matches rule or not. */
diff --git a/net/ipv6/netfilter/nf_conntrack_reasm.c b/net/ipv6/netfilter/nf_conntrack_reasm.c
index 578f3c1a16db..138a8b362706 100644
--- a/net/ipv6/netfilter/nf_conntrack_reasm.c
+++ b/net/ipv6/netfilter/nf_conntrack_reasm.c
@@ -363,7 +363,7 @@ nf_ct_frag6_reasm(struct nf_ct_frag6_queue *fq, struct net_device *dev)
        /* If the first fragment is fragmented itself, we split
         * it to two chunks: the first with data and paged part
         * and the second, holding only fragments. */
-        if (skb_has_frags(head)) {
+        if (skb_has_frag_list(head)) {
                struct sk_buff *clone;
                int i, plen = 0;
diff --git a/net/ipv6/protocol.c b/net/ipv6/protocol.c
index 1fa3468f0f32..9bb936ae2452 100644
--- a/net/ipv6/protocol.c
+++ b/net/ipv6/protocol.c
@@ -25,28 +25,14 @@
 #include <linux/spinlock.h>
 #include <net/protocol.h>
-const struct inet6_protocol *inet6_protos[MAX_INET_PROTOS];
+const struct inet6_protocol *inet6_protos[MAX_INET_PROTOS] __read_mostly;
-static DEFINE_SPINLOCK(inet6_proto_lock);
 int inet6_add_protocol(const struct inet6_protocol *prot, unsigned char protocol)
 {
-        int ret, hash = protocol & (MAX_INET_PROTOS - 1);
+        int hash = protocol & (MAX_INET_PROTOS - 1);
-        spin_lock_bh(&inet6_proto_lock);
-        if (inet6_protos[hash]) {
-                ret = -1;
-        } else {
-                inet6_protos[hash] = prot;
-                ret = 0;
-        }
-        spin_unlock_bh(&inet6_proto_lock);
-        return ret;
+        return !cmpxchg(&inet6_protos[hash], NULL, prot) ? 0 : -1;
 }
 EXPORT_SYMBOL(inet6_add_protocol);
 /*
@@ -57,20 +43,10 @@ int inet6_del_protocol(const struct inet6_protocol *prot, unsigned char protocol
 {
        int ret, hash = protocol & (MAX_INET_PROTOS - 1);
-        spin_lock_bh(&inet6_proto_lock);
+        ret = (cmpxchg(&inet6_protos[hash], prot, NULL) == prot) ? 0 : -1;
-        if (inet6_protos[hash] != prot) {
-                ret = -1;
-        } else {
-                inet6_protos[hash] = NULL;
-                ret = 0;
-        }
-        spin_unlock_bh(&inet6_proto_lock);
        synchronize_net();
        return ret;
 }
 EXPORT_SYMBOL(inet6_del_protocol);
diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c
index e677937a07fc..45e6efb7f171 100644
--- a/net/ipv6/raw.c
+++ b/net/ipv6/raw.c
@@ -764,7 +764,7 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
                        return -EINVAL;
                if (sin6->sin6_family && sin6->sin6_family != AF_INET6)
-                        return(-EAFNOSUPPORT);
+                        return -EAFNOSUPPORT;
                /* port is the proto value [0..255] carried in nexthdr */
                proto = ntohs(sin6->sin6_port);
@@ -772,10 +772,10 @@ static int rawv6_sendmsg(struct kiocb *iocb, struct sock *sk,
                if (!proto)
                        proto = inet->inet_num;
                else if (proto != inet->inet_num)
-                        return(-EINVAL);
+                        return -EINVAL;
                if (proto > 255)
-                        return(-EINVAL);
+                        return -EINVAL;
                daddr = &sin6->sin6_addr;
                if (np->sndflow) {
@@ -985,7 +985,7 @@ static int do_rawv6_setsockopt(struct sock *sk, int level, int optname,
                        /* You may get strange result with a positive odd offset;
                           RFC2292bis agrees with me. */
                        if (val > 0 && (val&1))
-                                return(-EINVAL);
+                                return -EINVAL;
                        if (val < 0) {
                                rp->checksum = 0;
                        } else {
@@ -997,7 +997,7 @@ static int do_rawv6_setsockopt(struct sock *sk, int level, int optname,
                        break;
                default:
-                        return(-ENOPROTOOPT);
+                        return -ENOPROTOOPT;
        }
 }
@@ -1190,7 +1190,7 @@ static int rawv6_init_sk(struct sock *sk)
        default:
                break;
        }
-        return(0);
+        return 0;
 }
 struct proto rawv6_prot = {
diff --git a/net/ipv6/reassembly.c b/net/ipv6/reassembly.c
index 64cfef1b0a4c..c7ba3149633f 100644
--- a/net/ipv6/reassembly.c
+++ b/net/ipv6/reassembly.c
@@ -458,7 +458,7 @@ static int ip6_frag_reasm(struct frag_queue *fq, struct sk_buff *prev,
        /* If the first fragment is fragmented itself, we split
         * it to two chunks: the first with data and paged part
         * and the second, holding only fragments. */
-        if (skb_has_frags(head)) {
+        if (skb_has_frag_list(head)) {
                struct sk_buff *clone;
                int i, plen = 0;
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index a275c6e1e25c..17e217933885 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -217,14 +217,14 @@ static void ip6_dst_ifdown(struct dst_entry *dst, struct net_device *dev,
 static __inline__ int rt6_check_expired(const struct rt6_info *rt)
 {
-        return (rt->rt6i_flags & RTF_EXPIRES &&
+        return (rt->rt6i_flags & RTF_EXPIRES) &&
-                time_after(jiffies, rt->rt6i_expires));
+                time_after(jiffies, rt->rt6i_expires);
 }
 static inline int rt6_need_strict(struct in6_addr *daddr)
 {
-        return (ipv6_addr_type(daddr) &
+        return ipv6_addr_type(daddr) &
-                (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK));
+                (IPV6_ADDR_MULTICAST | IPV6_ADDR_LINKLOCAL | IPV6_ADDR_LOOPBACK);
 }
 /*
@@ -440,7 +440,7 @@ static struct rt6_info *rt6_select(struct fib6_node *fn, int oif, int strict)
                  __func__, match);
        net = dev_net(rt0->rt6i_dev);
-        return (match ? match : net->ipv6.ip6_null_entry);
+        return match ? match : net->ipv6.ip6_null_entry;
 }
 #ifdef CONFIG_IPV6_ROUTE_INFO
@@ -859,7 +859,7 @@ int ip6_dst_blackhole(struct sock *sk, struct dst_entry **dstp, struct flowi *fl
        dst_release(*dstp);
        *dstp = new;
-        return (new ? 0 : -ENOMEM);
+        return new ? 0 : -ENOMEM;
 }
 EXPORT_SYMBOL_GPL(ip6_dst_blackhole);
@@ -1070,7 +1070,7 @@ static int ip6_dst_gc(struct dst_ops *ops)
                net->ipv6.ip6_rt_gc_expire = rt_gc_timeout>>1;
 out:
        net->ipv6.ip6_rt_gc_expire -= net->ipv6.ip6_rt_gc_expire>>rt_elasticity;
-        return (atomic_read(&ops->entries) > rt_max_size);
+        return atomic_read(&ops->entries) > rt_max_size;
 }
 /* Clean host part of a prefix. Not necessary in radix tree,
@@ -1169,6 +1169,8 @@ int ip6_route_add(struct fib6_config *cfg)
        if (addr_type & IPV6_ADDR_MULTICAST)
                rt->dst.input = ip6_mc_input;
+        else if (cfg->fc_flags & RTF_LOCAL)
+                rt->dst.input = ip6_input;
        else
                rt->dst.input = ip6_forward;
@@ -1190,7 +1192,8 @@ int ip6_route_add(struct fib6_config *cfg)
           they would result in kernel looping; promote them to reject routes
         */
        if ((cfg->fc_flags & RTF_REJECT) ||
-            (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK))) {
+            (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK)
+                                              && !(cfg->fc_flags&RTF_LOCAL))) {
                /* hold loopback dev/idev if we haven't done so. */
                if (dev != net->loopback_dev) {
                        if (dev) {
@@ -2102,6 +2105,9 @@ static int rtm_to_fib6_config(struct sk_buff *skb, struct nlmsghdr *nlh,
        if (rtm->rtm_type == RTN_UNREACHABLE)
                cfg->fc_flags |= RTF_REJECT;
+        if (rtm->rtm_type == RTN_LOCAL)
+                cfg->fc_flags |= RTF_LOCAL;
        cfg->fc_nlinfo.pid = NETLINK_CB(skb).pid;
        cfg->fc_nlinfo.nlh = nlh;
        cfg->fc_nlinfo.nl_net = sock_net(skb->sk);
@@ -2222,6 +2228,8 @@ static int rt6_fill_node(struct net *net,
        NLA_PUT_U32(skb, RTA_TABLE, table);
        if (rt->rt6i_flags&RTF_REJECT)
                rtm->rtm_type = RTN_UNREACHABLE;
+        else if (rt->rt6i_flags&RTF_LOCAL)
+                rtm->rtm_type = RTN_LOCAL;
        else if (rt->rt6i_dev && (rt->rt6i_dev->flags&IFF_LOOPBACK))
                rtm->rtm_type = RTN_LOCAL;
        else
diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c
index 4699cd3c3118..d7701782b639 100644
--- a/net/ipv6/sit.c
+++ b/net/ipv6/sit.c
@@ -63,36 +63,63 @@
 #define HASH_SIZE  16
 #define HASH(addr) (((__force u32)addr^((__force u32)addr>>4))&0xF)
-static void ipip6_tunnel_init(struct net_device *dev);
+static int ipip6_tunnel_init(struct net_device *dev);
 static void ipip6_tunnel_setup(struct net_device *dev);
+static void ipip6_dev_free(struct net_device *dev);
 static int sit_net_id __read_mostly;
 struct sit_net {
-        struct ip_tunnel *tunnels_r_l[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_r_l[HASH_SIZE];
-        struct ip_tunnel *tunnels_r[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_r[HASH_SIZE];
-        struct ip_tunnel *tunnels_l[HASH_SIZE];
+        struct ip_tunnel __rcu *tunnels_l[HASH_SIZE];
-        struct ip_tunnel *tunnels_wc[1];
+        struct ip_tunnel __rcu *tunnels_wc[1];
-        struct ip_tunnel **tunnels[4];
+        struct ip_tunnel __rcu **tunnels[4];
        struct net_device *fb_tunnel_dev;
 };
 /*
- * Locking : hash tables are protected by RCU and a spinlock
+ * Locking : hash tables are protected by RCU and RTNL
 */
-static DEFINE_SPINLOCK(ipip6_lock);
 #define for_each_ip_tunnel_rcu(start) \
        for (t = rcu_dereference(start); t; t = rcu_dereference(t->next))
+/* often modified stats are per cpu, other are shared (netdev->stats) */
+struct pcpu_tstats {
+        unsigned long   rx_packets;
+        unsigned long   rx_bytes;
+        unsigned long   tx_packets;
+        unsigned long   tx_bytes;
+};
+static struct net_device_stats *ipip6_get_stats(struct net_device *dev)
+{
+        struct pcpu_tstats sum = { 0 };
+        int i;
+        for_each_possible_cpu(i) {
+                const struct pcpu_tstats *tstats = per_cpu_ptr(dev->tstats, i);
+                sum.rx_packets += tstats->rx_packets;
+                sum.rx_bytes   += tstats->rx_bytes;
+                sum.tx_packets += tstats->tx_packets;
+                sum.tx_bytes   += tstats->tx_bytes;
+        }
+        dev->stats.rx_packets = sum.rx_packets;
+        dev->stats.rx_bytes   = sum.rx_bytes;
+        dev->stats.tx_packets = sum.tx_packets;
+        dev->stats.tx_bytes   = sum.tx_bytes;
+        return &dev->stats;
+}
 /*
 * Must be invoked with rcu_read_lock
 */
 static struct ip_tunnel * ipip6_tunnel_lookup(struct net *net,
                struct net_device *dev, __be32 remote, __be32 local)
 {
-        unsigned h0 = HASH(remote);
+        unsigned int h0 = HASH(remote);
-        unsigned h1 = HASH(local);
+        unsigned int h1 = HASH(local);
        struct ip_tunnel *t;
        struct sit_net *sitn = net_generic(net, sit_net_id);
@@ -121,12 +148,12 @@ static struct ip_tunnel * ipip6_tunnel_lookup(struct net *net,
        return NULL;
 }
-static struct ip_tunnel **__ipip6_bucket(struct sit_net *sitn,
+static struct ip_tunnel __rcu **__ipip6_bucket(struct sit_net *sitn,
                struct ip_tunnel_parm *parms)
 {
        __be32 remote = parms->iph.daddr;
        __be32 local = parms->iph.saddr;
-        unsigned h = 0;
+        unsigned int h = 0;
        int prio = 0;
        if (remote) {
@@ -140,7 +167,7 @@ static struct ip_tunnel **__ipip6_bucket(struct sit_net *sitn,
        return &sitn->tunnels[prio][h];
 }
-static inline struct ip_tunnel **ipip6_bucket(struct sit_net *sitn,
+static inline struct ip_tunnel __rcu **ipip6_bucket(struct sit_net *sitn,
                struct ip_tunnel *t)
 {
        return __ipip6_bucket(sitn, &t->parms);
@@ -148,13 +175,14 @@ static inline struct ip_tunnel **ipip6_bucket(struct sit_net *sitn,
 static void ipip6_tunnel_unlink(struct sit_net *sitn, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp;
+        struct ip_tunnel __rcu **tp;
+        struct ip_tunnel *iter;
-        for (tp = ipip6_bucket(sitn, t); *tp; tp = &(*tp)->next) {
-                if (t == *tp) {
+        for (tp = ipip6_bucket(sitn, t);
-                        spin_lock_bh(&ipip6_lock);
+             (iter = rtnl_dereference(*tp)) != NULL;
-                        *tp = t->next;
+             tp = &iter->next) {
-                        spin_unlock_bh(&ipip6_lock);
+                if (t == iter) {
+                        rcu_assign_pointer(*tp, t->next);
                        break;
                }
        }
@@ -162,12 +190,10 @@ static void ipip6_tunnel_unlink(struct sit_net *sitn, struct ip_tunnel *t)
 static void ipip6_tunnel_link(struct sit_net *sitn, struct ip_tunnel *t)
 {
-        struct ip_tunnel **tp = ipip6_bucket(sitn, t);
+        struct ip_tunnel __rcu **tp = ipip6_bucket(sitn, t);
-        spin_lock_bh(&ipip6_lock);
+        rcu_assign_pointer(t->next, rtnl_dereference(*tp));
-        t->next = *tp;
        rcu_assign_pointer(*tp, t);
-        spin_unlock_bh(&ipip6_lock);
 }
 static void ipip6_tunnel_clone_6rd(struct net_device *dev, struct sit_net *sitn)
@@ -187,17 +213,20 @@ static void ipip6_tunnel_clone_6rd(struct net_device *dev, struct sit_net *sitn)
 #endif
 }
-static struct ip_tunnel * ipip6_tunnel_locate(struct net *net,
+static struct ip_tunnel *ipip6_tunnel_locate(struct net *net,
                struct ip_tunnel_parm *parms, int create)
 {
        __be32 remote = parms->iph.daddr;
        __be32 local = parms->iph.saddr;
-        struct ip_tunnel *t, **tp, *nt;
+        struct ip_tunnel *t, *nt;
+        struct ip_tunnel __rcu **tp;
        struct net_device *dev;
        char name[IFNAMSIZ];
        struct sit_net *sitn = net_generic(net, sit_net_id);
-        for (tp = __ipip6_bucket(sitn, parms); (t = *tp) != NULL; tp = &t->next) {
+        for (tp = __ipip6_bucket(sitn, parms);
+            (t = rtnl_dereference(*tp)) != NULL;
+             tp = &t->next) {
                if (local == t->parms.iph.saddr &&
                    remote == t->parms.iph.daddr &&
                    parms->link == t->parms.link) {
@@ -213,7 +242,7 @@ static struct ip_tunnel * ipip6_tunnel_locate(struct net *net,
        if (parms->name[0])
                strlcpy(name, parms->name, IFNAMSIZ);
        else
-                sprintf(name, "sit%%d");
+                strcpy(name, "sit%d");
        dev = alloc_netdev(sizeof(*t), name, ipip6_tunnel_setup);
        if (dev == NULL)
@@ -229,7 +258,8 @@ static struct ip_tunnel * ipip6_tunnel_locate(struct net *net,
        nt = netdev_priv(dev);
        nt->parms = *parms;
-        ipip6_tunnel_init(dev);
+        if (ipip6_tunnel_init(dev) < 0)
+                goto failed_free;
        ipip6_tunnel_clone_6rd(dev, sitn);
        if (parms->i_flags & SIT_ISATAP)
@@ -244,7 +274,7 @@ static struct ip_tunnel * ipip6_tunnel_locate(struct net *net,
        return nt;
 failed_free:
-        free_netdev(dev);
+        ipip6_dev_free(dev);
 failed:
        return NULL;
 }
@@ -340,7 +370,7 @@ ipip6_tunnel_add_prl(struct ip_tunnel *t, struct ip_tunnel_prl *a, int chg)
        ASSERT_RTNL();
-        for (p = t->prl; p; p = p->next) {
+        for (p = rtnl_dereference(t->prl); p; p = rtnl_dereference(p->next)) {
                if (p->addr == a->addr) {
                        if (chg) {
                                p->flags = a->flags;
@@ -451,15 +481,12 @@ static void ipip6_tunnel_uninit(struct net_device *dev)
        struct sit_net *sitn = net_generic(net, sit_net_id);
        if (dev == sitn->fb_tunnel_dev) {
-                spin_lock_bh(&ipip6_lock);
+                rcu_assign_pointer(sitn->tunnels_wc[0], NULL);
-                sitn->tunnels_wc[0] = NULL;
-                spin_unlock_bh(&ipip6_lock);
-                dev_put(dev);
        } else {
                ipip6_tunnel_unlink(sitn, netdev_priv(dev));
                ipip6_tunnel_del_prl(netdev_priv(dev), NULL);
-                dev_put(dev);
        }
+        dev_put(dev);
 }
@@ -548,6 +575,8 @@ static int ipip6_rcv(struct sk_buff *skb)
        tunnel = ipip6_tunnel_lookup(dev_net(skb->dev), skb->dev,
                                     iph->saddr, iph->daddr);
        if (tunnel != NULL) {
+                struct pcpu_tstats *tstats;
                secpath_reset(skb);
                skb->mac_header = skb->network_header;
                skb_reset_network_header(skb);
@@ -563,10 +592,17 @@ static int ipip6_rcv(struct sk_buff *skb)
                        return 0;
                }
-                skb_tunnel_rx(skb, tunnel->dev);
+                tstats = this_cpu_ptr(tunnel->dev->tstats);
+                tstats->rx_packets++;
+                tstats->rx_bytes += skb->len;
+                __skb_tunnel_rx(skb, tunnel->dev);
                ipip6_ecn_decapsulate(iph, skb);
-                netif_rx(skb);
+                if (netif_rx(skb) == NET_RX_DROP)
+                        tunnel->dev->stats.rx_dropped++;
                rcu_read_unlock();
                return 0;
        }
@@ -590,7 +626,7 @@ __be32 try_6rd(struct in6_addr *v6dst, struct ip_tunnel *tunnel)
 #ifdef CONFIG_IPV6_SIT_6RD
        if (ipv6_prefix_equal(v6dst, &tunnel->ip6rd.prefix,
                              tunnel->ip6rd.prefixlen)) {
-                unsigned pbw0, pbi0;
+                unsigned int pbw0, pbi0;
                int pbi1;
                u32 d;
@@ -625,14 +661,13 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
                                     struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
-        struct net_device_stats *stats = &dev->stats;
+        struct pcpu_tstats *tstats;
-        struct netdev_queue *txq = netdev_get_tx_queue(dev, 0);
        struct iphdr  *tiph = &tunnel->parms.iph;
        struct ipv6hdr *iph6 = ipv6_hdr(skb);
        u8     tos = tunnel->parms.iph.tos;
        __be16 df = tiph->frag_off;
        struct rtable *rt;                      /* Route to the other host */
-        struct net_device *tdev;                        /* Device to other host */
+        struct net_device *tdev;                /* Device to other host */
        struct iphdr  *iph;                     /* Our new IP header */
        unsigned int max_headroom;              /* The extra header space needed */
        __be32 dst = tiph->daddr;
@@ -703,20 +738,20 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
                                    .oif = tunnel->parms.link,
                                    .proto = IPPROTO_IPV6 };
                if (ip_route_output_key(dev_net(dev), &rt, &fl)) {
-                        stats->tx_carrier_errors++;
+                        dev->stats.tx_carrier_errors++;
                        goto tx_error_icmp;
                }
        }
        if (rt->rt_type != RTN_UNICAST) {
                ip_rt_put(rt);
-                stats->tx_carrier_errors++;
+                dev->stats.tx_carrier_errors++;
                goto tx_error_icmp;
        }
        tdev = rt->dst.dev;
        if (tdev == dev) {
                ip_rt_put(rt);
-                stats->collisions++;
+                dev->stats.collisions++;
                goto tx_error;
        }
@@ -724,7 +759,7 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
                mtu = dst_mtu(&rt->dst) - sizeof(struct iphdr);
                if (mtu < 68) {
-                        stats->collisions++;
+                        dev->stats.collisions++;
                        ip_rt_put(rt);
                        goto tx_error;
                }
@@ -763,7 +798,7 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
                struct sk_buff *new_skb = skb_realloc_headroom(skb, max_headroom);
                if (!new_skb) {
                        ip_rt_put(rt);
-                        txq->tx_dropped++;
+                        dev->stats.tx_dropped++;
                        dev_kfree_skb(skb);
                        return NETDEV_TX_OK;
                }
@@ -799,14 +834,14 @@ static netdev_tx_t ipip6_tunnel_xmit(struct sk_buff *skb,
                iph->ttl        =       iph6->hop_limit;
        nf_reset(skb);
+        tstats = this_cpu_ptr(dev->tstats);
-        IPTUNNEL_XMIT();
+        __IPTUNNEL_XMIT(tstats, &dev->stats);
        return NETDEV_TX_OK;
 tx_error_icmp:
        dst_link_failure(skb);
 tx_error:
-        stats->tx_errors++;
+        dev->stats.tx_errors++;
        dev_kfree_skb(skb);
        return NETDEV_TX_OK;
 }
@@ -1083,12 +1118,19 @@ static const struct net_device_ops ipip6_netdev_ops = {
        .ndo_start_xmit = ipip6_tunnel_xmit,
        .ndo_do_ioctl   = ipip6_tunnel_ioctl,
        .ndo_change_mtu = ipip6_tunnel_change_mtu,
+        .ndo_get_stats  = ipip6_get_stats,
 };
+static void ipip6_dev_free(struct net_device *dev)
+{
+        free_percpu(dev->tstats);
+        free_netdev(dev);
+}
 static void ipip6_tunnel_setup(struct net_device *dev)
 {
        dev->netdev_ops         = &ipip6_netdev_ops;
-        dev->destructor         = free_netdev;
+        dev->destructor         = ipip6_dev_free;
        dev->type               = ARPHRD_SIT;
        dev->hard_header_len    = LL_MAX_HEADER + sizeof(struct iphdr);
@@ -1098,9 +1140,10 @@ static void ipip6_tunnel_setup(struct net_device *dev)
        dev->iflink             = 0;
        dev->addr_len           = 4;
        dev->features           |= NETIF_F_NETNS_LOCAL;
+        dev->features           |= NETIF_F_LLTX;
 }
-static void ipip6_tunnel_init(struct net_device *dev)
+static int ipip6_tunnel_init(struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
@@ -1111,9 +1154,14 @@ static void ipip6_tunnel_init(struct net_device *dev)
        memcpy(dev->broadcast, &tunnel->parms.iph.daddr, 4);
        ipip6_tunnel_bind_dev(dev);
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
+        return 0;
 }
-static void __net_init ipip6_fb_tunnel_init(struct net_device *dev)
+static int __net_init ipip6_fb_tunnel_init(struct net_device *dev)
 {
        struct ip_tunnel *tunnel = netdev_priv(dev);
        struct iphdr *iph = &tunnel->parms.iph;
@@ -1128,11 +1176,15 @@ static void __net_init ipip6_fb_tunnel_init(struct net_device *dev)
        iph->ihl                = 5;
        iph->ttl                = 64;
+        dev->tstats = alloc_percpu(struct pcpu_tstats);
+        if (!dev->tstats)
+                return -ENOMEM;
        dev_hold(dev);
        sitn->tunnels_wc[0]     = tunnel;
+        return 0;
 }
-static struct xfrm_tunnel sit_handler = {
+static struct xfrm_tunnel sit_handler __read_mostly = {
        .handler        =       ipip6_rcv,
        .err_handler    =       ipip6_err,
        .priority       =       1,
@@ -1173,7 +1225,10 @@ static int __net_init sit_init_net(struct net *net)
        }
        dev_net_set(sitn->fb_tunnel_dev, net);
-        ipip6_fb_tunnel_init(sitn->fb_tunnel_dev);
+        err = ipip6_fb_tunnel_init(sitn->fb_tunnel_dev);
+        if (err)
+                goto err_dev_free;
        ipip6_tunnel_clone_6rd(sitn->fb_tunnel_dev, sitn);
        if ((err = register_netdev(sitn->fb_tunnel_dev)))
@@ -1183,7 +1238,8 @@ static int __net_init sit_init_net(struct net *net)
 err_reg_dev:
        dev_put(sitn->fb_tunnel_dev);
-        free_netdev(sitn->fb_tunnel_dev);
+err_dev_free:
+        ipip6_dev_free(sitn->fb_tunnel_dev);
 err_alloc_dev:
        return err;
 }
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index fe6d40418c0b..8d93f6d81979 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -139,7 +139,7 @@ static int tcp_v6_connect(struct sock *sk, struct sockaddr *uaddr,
                return -EINVAL;
        if (usin->sin6_family != AF_INET6)
-                return(-EAFNOSUPPORT);
+                return -EAFNOSUPPORT;
        memset(&fl, 0, sizeof(fl));
diff --git a/net/ipv6/tunnel6.c b/net/ipv6/tunnel6.c
index fc3c86a47452..d9864725d0c6 100644
--- a/net/ipv6/tunnel6.c
+++ b/net/ipv6/tunnel6.c
@@ -30,8 +30,8 @@
 #include <net/protocol.h>
 #include <net/xfrm.h>
-static struct xfrm6_tunnel *tunnel6_handlers;
+static struct xfrm6_tunnel *tunnel6_handlers __read_mostly;
-static struct xfrm6_tunnel *tunnel46_handlers;
+static struct xfrm6_tunnel *tunnel46_handlers __read_mostly;
 static DEFINE_MUTEX(tunnel6_mutex);
 int xfrm6_tunnel_register(struct xfrm6_tunnel *handler, unsigned short family)
@@ -51,7 +51,7 @@ int xfrm6_tunnel_register(struct xfrm6_tunnel *handler, unsigned short family)
        }
        handler->next = *pprev;
-        *pprev = handler;
+        rcu_assign_pointer(*pprev, handler);
        ret = 0;
@@ -88,6 +88,11 @@ int xfrm6_tunnel_deregister(struct xfrm6_tunnel *handler, unsigned short family)
 EXPORT_SYMBOL(xfrm6_tunnel_deregister);
+#define for_each_tunnel_rcu(head, handler)              \
+        for (handler = rcu_dereference(head);           \
+             handler != NULL;                           \
+             handler = rcu_dereference(handler->next))  \
 static int tunnel6_rcv(struct sk_buff *skb)
 {
        struct xfrm6_tunnel *handler;
@@ -95,7 +100,7 @@ static int tunnel6_rcv(struct sk_buff *skb)
        if (!pskb_may_pull(skb, sizeof(struct ipv6hdr)))
                goto drop;
-        for (handler = tunnel6_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel6_handlers, handler)
                if (!handler->handler(skb))
                        return 0;
@@ -113,7 +118,7 @@ static int tunnel46_rcv(struct sk_buff *skb)
        if (!pskb_may_pull(skb, sizeof(struct iphdr)))
                goto drop;
-        for (handler = tunnel46_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel46_handlers, handler)
                if (!handler->handler(skb))
                        return 0;
@@ -129,7 +134,7 @@ static void tunnel6_err(struct sk_buff *skb, struct inet6_skb_parm *opt,
 {
        struct xfrm6_tunnel *handler;
-        for (handler = tunnel6_handlers; handler; handler = handler->next)
+        for_each_tunnel_rcu(tunnel6_handlers, handler)
                if (!handler->err_handler(skb, opt, type, code, offset, info))
                        break;
 }
diff --git a/net/ipv6/xfrm6_policy.c b/net/ipv6/xfrm6_policy.c
index 6baeabbbca82..39676eac3a37 100644
--- a/net/ipv6/xfrm6_policy.c
+++ b/net/ipv6/xfrm6_policy.c
@@ -199,7 +199,7 @@ static inline int xfrm6_garbage_collect(struct dst_ops *ops)
        struct net *net = container_of(ops, struct net, xfrm.xfrm6_dst_ops);
        xfrm6_policy_afinfo.garbage_collect(net);
-        return (atomic_read(&ops->entries) > ops->gc_thresh * 2);
+        return atomic_read(&ops->entries) > ops->gc_thresh * 2;
 }
 static void xfrm6_update_pmtu(struct dst_entry *dst, u32 mtu)
diff --git a/net/ipv6/xfrm6_tunnel.c b/net/ipv6/xfrm6_tunnel.c
index 2ce3a8278f26..ac7584b946a5 100644
--- a/net/ipv6/xfrm6_tunnel.c
+++ b/net/ipv6/xfrm6_tunnel.c
@@ -317,13 +317,13 @@ static const struct xfrm_type xfrm6_tunnel_type = {
        .output         = xfrm6_tunnel_output,
 };
-static struct xfrm6_tunnel xfrm6_tunnel_handler = {
+static struct xfrm6_tunnel xfrm6_tunnel_handler __read_mostly = {
        .handler        = xfrm6_tunnel_rcv,
        .err_handler    = xfrm6_tunnel_err,
        .priority       = 2,
 };
-static struct xfrm6_tunnel xfrm46_tunnel_handler = {
+static struct xfrm6_tunnel xfrm46_tunnel_handler __read_mostly = {
        .handler        = xfrm6_tunnel_rcv,
        .err_handler    = xfrm6_tunnel_err,
        .priority       = 2,
diff --git a/net/irda/af_irda.c b/net/irda/af_irda.c
index fd55b5135de5..bf3635129b17 100644
--- a/net/irda/af_irda.c
+++ b/net/irda/af_irda.c
@@ -573,9 +573,9 @@ static int irda_find_lsap_sel(struct irda_sock *self, char *name)
                /* Requested object/attribute doesn't exist */
                if((self->errno == IAS_CLASS_UNKNOWN) ||
                   (self->errno == IAS_ATTRIB_UNKNOWN))
-                        return (-EADDRNOTAVAIL);
+                        return -EADDRNOTAVAIL;
                else
-                        return (-EHOSTUNREACH);
+                        return -EHOSTUNREACH;
        }
        /* Get the remote TSAP selector */
@@ -663,7 +663,7 @@ static int irda_discover_daddr_and_lsap_sel(struct irda_sock *self, char *name)
                                           __func__, name);
                                self->daddr = DEV_ADDR_ANY;
                                kfree(discoveries);
-                                return(-ENOTUNIQ);
+                                return -ENOTUNIQ;
                        }
                        /* First time we found that one, save it ! */
                        daddr = self->daddr;
@@ -677,7 +677,7 @@ static int irda_discover_daddr_and_lsap_sel(struct irda_sock *self, char *name)
                        IRDA_DEBUG(0, "%s(), unexpected IAS query failure\n", __func__);
                        self->daddr = DEV_ADDR_ANY;
                        kfree(discoveries);
-                        return(-EHOSTUNREACH);
+                        return -EHOSTUNREACH;
                        break;
                }
        }
@@ -689,7 +689,7 @@ static int irda_discover_daddr_and_lsap_sel(struct irda_sock *self, char *name)
                IRDA_DEBUG(1, "%s(), cannot discover service ''%s'' in any device !!!\n",
                           __func__, name);
                self->daddr = DEV_ADDR_ANY;
-                return(-EADDRNOTAVAIL);
+                return -EADDRNOTAVAIL;
        }
        /* Revert back to discovered device & service */
@@ -2465,9 +2465,9 @@ bed:
                        /* Requested object/attribute doesn't exist */
                        if((self->errno == IAS_CLASS_UNKNOWN) ||
                           (self->errno == IAS_ATTRIB_UNKNOWN))
-                                return (-EADDRNOTAVAIL);
+                                return -EADDRNOTAVAIL;
                        else
-                                return (-EHOSTUNREACH);
+                                return -EHOSTUNREACH;
                }
                /* Translate from internal to user structure */
diff --git a/net/irda/discovery.c b/net/irda/discovery.c
index c1c8ae939126..36c3f037f172 100644
--- a/net/irda/discovery.c
+++ b/net/irda/discovery.c
@@ -315,7 +315,7 @@ struct irda_device_info *irlmp_copy_discoveries(hashbin_t *log, int *pn,
        /* Get the actual number of device in the buffer and return */
        *pn = i;
-        return(buffer);
+        return buffer;
 }
 #ifdef CONFIG_PROC_FS
diff --git a/net/irda/ircomm/ircomm_tty.c b/net/irda/ircomm/ircomm_tty.c
index faa82ca2dfdc..a39cca8331df 100644
--- a/net/irda/ircomm/ircomm_tty.c
+++ b/net/irda/ircomm/ircomm_tty.c
@@ -449,8 +449,8 @@ static int ircomm_tty_open(struct tty_struct *tty, struct file *filp)
                }
 #ifdef SERIAL_DO_RESTART
-                return ((self->flags & ASYNC_HUP_NOTIFY) ?
+                return (self->flags & ASYNC_HUP_NOTIFY) ?
-                        -EAGAIN : -ERESTARTSYS);
+                        -EAGAIN : -ERESTARTSYS;
 #else
                return -EAGAIN;
 #endif
diff --git a/net/irda/irlan/irlan_eth.c b/net/irda/irlan/irlan_eth.c
index 5bb8353105cc..8ee1ff6c742f 100644
--- a/net/irda/irlan/irlan_eth.c
+++ b/net/irda/irlan/irlan_eth.c
@@ -45,13 +45,11 @@ static int  irlan_eth_close(struct net_device *dev);
 static netdev_tx_t  irlan_eth_xmit(struct sk_buff *skb,
                                         struct net_device *dev);
 static void irlan_eth_set_multicast_list( struct net_device *dev);
-static struct net_device_stats *irlan_eth_get_stats(struct net_device *dev);
 static const struct net_device_ops irlan_eth_netdev_ops = {
        .ndo_open               = irlan_eth_open,
        .ndo_stop               = irlan_eth_close,
        .ndo_start_xmit         = irlan_eth_xmit,
-        .ndo_get_stats          = irlan_eth_get_stats,
        .ndo_set_multicast_list = irlan_eth_set_multicast_list,
        .ndo_change_mtu         = eth_change_mtu,
        .ndo_validate_addr      = eth_validate_addr,
@@ -208,10 +206,10 @@ static netdev_tx_t irlan_eth_xmit(struct sk_buff *skb,
                 * tried :-) DB
                 */
                /* irttp_data_request already free the packet */
-                self->stats.tx_dropped++;
+                dev->stats.tx_dropped++;
        } else {
-                self->stats.tx_packets++;
+                dev->stats.tx_packets++;
-                self->stats.tx_bytes += len;
+                dev->stats.tx_bytes += len;
        }
        return NETDEV_TX_OK;
@@ -226,15 +224,16 @@ static netdev_tx_t irlan_eth_xmit(struct sk_buff *skb,
 int irlan_eth_receive(void *instance, void *sap, struct sk_buff *skb)
 {
        struct irlan_cb *self = instance;
+        struct net_device *dev = self->dev;
        if (skb == NULL) {
-                ++self->stats.rx_dropped;
+                dev->stats.rx_dropped++;
                return 0;
        }
        if (skb->len < ETH_HLEN) {
                IRDA_DEBUG(0, "%s() : IrLAN frame too short (%d)\n",
                           __func__, skb->len);
-                ++self->stats.rx_dropped;
+                dev->stats.rx_dropped++;
                dev_kfree_skb(skb);
                return 0;
        }
@@ -244,10 +243,10 @@ int irlan_eth_receive(void *instance, void *sap, struct sk_buff *skb)
         * might have been previously set by the low level IrDA network
         * device driver
         */
-        skb->protocol = eth_type_trans(skb, self->dev); /* Remove eth header */
+        skb->protocol = eth_type_trans(skb, dev); /* Remove eth header */
-        self->stats.rx_packets++;
+        dev->stats.rx_packets++;
-        self->stats.rx_bytes += skb->len;
+        dev->stats.rx_bytes += skb->len;
        netif_rx(skb);   /* Eat it! */
@@ -348,16 +347,3 @@ static void irlan_eth_set_multicast_list(struct net_device *dev)
        else
                irlan_set_broadcast_filter(self, FALSE);
 }
-/*
- * Function irlan_get_stats (dev)
- *
- *    Get the current statistics for this device
- *
- */
-static struct net_device_stats *irlan_eth_get_stats(struct net_device *dev)
-{
-        struct irlan_cb *self = netdev_priv(dev);
-        return &self->stats;
-}
diff --git a/net/irda/irlan/irlan_event.c b/net/irda/irlan/irlan_event.c
index cbcb4eb54037..43f16040a6fe 100644
--- a/net/irda/irlan/irlan_event.c
+++ b/net/irda/irlan/irlan_event.c
@@ -24,7 +24,7 @@
 #include <net/irda/irlan_event.h>
-char *irlan_state[] = {
+const char * const irlan_state[] = {
        "IRLAN_IDLE",
        "IRLAN_QUERY",
        "IRLAN_CONN",
diff --git a/net/irda/irlmp.c b/net/irda/irlmp.c
index 0e7d8bde145d..6115a44c0a24 100644
--- a/net/irda/irlmp.c
+++ b/net/irda/irlmp.c
@@ -939,7 +939,7 @@ struct irda_device_info *irlmp_get_discoveries(int *pn, __u16 mask, int nslots)
        }
        /* Return current cached discovery log */
-        return(irlmp_copy_discoveries(irlmp->cachelog, pn, mask, TRUE));
+        return irlmp_copy_discoveries(irlmp->cachelog, pn, mask, TRUE);
 }
 EXPORT_SYMBOL(irlmp_get_discoveries);
diff --git a/net/irda/irlmp_frame.c b/net/irda/irlmp_frame.c
index 3750884094da..062e63b1c5c4 100644
--- a/net/irda/irlmp_frame.c
+++ b/net/irda/irlmp_frame.c
@@ -448,7 +448,7 @@ static struct lsap_cb *irlmp_find_lsap(struct lap_cb *self, __u8 dlsap_sel,
            (self->cache.slsap_sel == slsap_sel) &&
            (self->cache.dlsap_sel == dlsap_sel))
        {
-                return (self->cache.lsap);
+                return self->cache.lsap;
        }
 #endif
diff --git a/net/irda/irnet/irnet_irda.c b/net/irda/irnet/irnet_irda.c
index e98e40d76f4f..7f17a8020e8a 100644
--- a/net/irda/irnet/irnet_irda.c
+++ b/net/irda/irnet/irnet_irda.c
@@ -238,7 +238,7 @@ irnet_ias_to_tsap(irnet_socket *	self,
  DEXIT(IRDA_SR_TRACE, "\n");
  /* Return the TSAP */
-  return(dtsap_sel);
+  return dtsap_sel;
 }
 /*------------------------------------------------------------------*/
@@ -301,7 +301,7 @@ irnet_connect_tsap(irnet_socket *	self)
    {
      clear_bit(0, &self->ttp_connect);
      DERROR(IRDA_SR_ERROR, "connect aborted!\n");
-      return(err);
+      return err;
    }
  /* Connect to remote device */
@@ -312,7 +312,7 @@ irnet_connect_tsap(irnet_socket *	self)
    {
      clear_bit(0, &self->ttp_connect);
      DERROR(IRDA_SR_ERROR, "connect aborted!\n");
-      return(err);
+      return err;
    }
  /* The above call is non-blocking.
@@ -321,7 +321,7 @@ irnet_connect_tsap(irnet_socket *	self)
   * See you there ;-) */
  DEXIT(IRDA_SR_TRACE, "\n");
-  return(err);
+  return err;
 }
 /*------------------------------------------------------------------*/
@@ -362,10 +362,10 @@ irnet_discover_next_daddr(irnet_socket *	self)
      /* The above request is non-blocking.
       * After a while, IrDA will call us back in irnet_discovervalue_confirm()
       * We will then call irnet_ias_to_tsap() and come back here again... */
-      return(0);
+      return 0;
    }
  else
-    return(1);
+    return 1;
 }
 /*------------------------------------------------------------------*/
@@ -436,7 +436,7 @@ irnet_discover_daddr_and_lsap_sel(irnet_socket *	self)
  /* Follow me in irnet_discovervalue_confirm() */
  DEXIT(IRDA_SR_TRACE, "\n");
-  return(0);
+  return 0;
 }
 /*------------------------------------------------------------------*/
@@ -485,7 +485,7 @@ irnet_dname_to_daddr(irnet_socket *	self)
  /* No luck ! */
  DEBUG(IRDA_SR_INFO, "cannot discover device ``%s'' !!!\n", self->rname);
  kfree(discoveries);
-  return(-EADDRNOTAVAIL);
+  return -EADDRNOTAVAIL;
 }
@@ -527,7 +527,7 @@ irda_irnet_create(irnet_socket *	self)
  INIT_WORK(&self->disconnect_work, irnet_ppp_disconnect);
  DEXIT(IRDA_SOCK_TRACE, "\n");
-  return(0);
+  return 0;
 }
 /*------------------------------------------------------------------*/
@@ -601,7 +601,7 @@ irda_irnet_connect(irnet_socket *	self)
   * We will finish the connection procedure in irnet_connect_tsap().
   */
  DEXIT(IRDA_SOCK_TRACE, "\n");
-  return(0);
+  return 0;
 }
 /*------------------------------------------------------------------*/
@@ -733,7 +733,7 @@ irnet_daddr_to_dname(irnet_socket *	self)
  /* No luck ! */
  DEXIT(IRDA_SERV_INFO, ": cannot discover device 0x%08x !!!\n", self->daddr);
  kfree(discoveries);
-  return(-EADDRNOTAVAIL);
+  return -EADDRNOTAVAIL;
 }
 /*------------------------------------------------------------------*/
diff --git a/net/irda/irnet/irnet_ppp.c b/net/irda/irnet/irnet_ppp.c
index dfe7b38dd4af..69f1fa64994e 100644
--- a/net/irda/irnet/irnet_ppp.c
+++ b/net/irda/irnet/irnet_ppp.c
@@ -166,7 +166,7 @@ irnet_ctrl_write(irnet_socket *	ap,
    }
  /* Success : we have parsed all commands successfully */
-  return(count);
+  return count;
 }
 #ifdef INITIAL_DISCOVERY
@@ -300,7 +300,7 @@ irnet_ctrl_read(irnet_socket *	ap,
        }
      DEXIT(CTRL_TRACE, "\n");
-      return(strlen(event));
+      return strlen(event);
    }
 #endif /* INITIAL_DISCOVERY */
@@ -409,7 +409,7 @@ irnet_ctrl_read(irnet_socket *	ap,
    }
  DEXIT(CTRL_TRACE, "\n");
-  return(strlen(event));
+  return strlen(event);
 }
 /*------------------------------------------------------------------*/
@@ -623,7 +623,7 @@ dev_irnet_poll(struct file *	file,
    mask |= irnet_ctrl_poll(ap, file, wait);
  DEXIT(FS_TRACE, " - mask=0x%X\n", mask);
-  return(mask);
+  return mask;
 }
 /*------------------------------------------------------------------*/
diff --git a/net/irda/irnet/irnet_ppp.h b/net/irda/irnet/irnet_ppp.h
index b5df2418f90c..940225866da0 100644
--- a/net/irda/irnet/irnet_ppp.h
+++ b/net/irda/irnet/irnet_ppp.h
@@ -103,7 +103,8 @@ static const struct file_operations irnet_device_fops =
        .poll           = dev_irnet_poll,
        .unlocked_ioctl = dev_irnet_ioctl,
        .open           = dev_irnet_open,
-        .release        = dev_irnet_close
+        .release        = dev_irnet_close,
+        .llseek         = noop_llseek,
  /* Also : llseek, readdir, mmap, flush, fsync, fasync, lock, readv, writev */
 };
diff --git a/net/key/af_key.c b/net/key/af_key.c
index 43040e97c474..d87c22df6f1e 100644
--- a/net/key/af_key.c
+++ b/net/key/af_key.c
@@ -565,12 +565,12 @@ pfkey_proto2satype(uint16_t proto)
 static uint8_t pfkey_proto_to_xfrm(uint8_t proto)
 {
-        return (proto == IPSEC_PROTO_ANY ? 0 : proto);
+        return proto == IPSEC_PROTO_ANY ? 0 : proto;
 }
 static uint8_t pfkey_proto_from_xfrm(uint8_t proto)
 {
-        return (proto ? proto : IPSEC_PROTO_ANY);
+        return proto ? proto : IPSEC_PROTO_ANY;
 }
 static inline int pfkey_sockaddr_len(sa_family_t family)
diff --git a/net/l2tp/l2tp_eth.c b/net/l2tp/l2tp_eth.c
index 1ae697681bc7..8d9ce0accc98 100644
--- a/net/l2tp/l2tp_eth.c
+++ b/net/l2tp/l2tp_eth.c
@@ -144,7 +144,6 @@ static void l2tp_eth_dev_recv(struct l2tp_session *session, struct sk_buff *skb,
        nf_reset(skb);
        if (dev_forward_skb(dev, skb) == NET_RX_SUCCESS) {
-                dev->last_rx = jiffies;
                dev->stats.rx_packets++;
                dev->stats.rx_bytes += data_len;
        } else
diff --git a/net/l2tp/l2tp_ppp.c b/net/l2tp/l2tp_ppp.c
index ff954b3e94b6..39a21d0c61c4 100644
--- a/net/l2tp/l2tp_ppp.c
+++ b/net/l2tp/l2tp_ppp.c
@@ -1768,7 +1768,7 @@ static const struct proto_ops pppol2tp_ops = {
        .ioctl          = pppox_ioctl,
 };
-static struct pppox_proto pppol2tp_proto = {
+static const struct pppox_proto pppol2tp_proto = {
        .create         = pppol2tp_create,
        .ioctl          = pppol2tp_ioctl
 };
diff --git a/net/mac80211/aes_ccm.c b/net/mac80211/aes_ccm.c
index a87cb3ba2df6..d2b03e0851ef 100644
--- a/net/mac80211/aes_ccm.c
+++ b/net/mac80211/aes_ccm.c
@@ -138,10 +138,8 @@ struct crypto_cipher *ieee80211_aes_key_setup_encrypt(const u8 key[])
        struct crypto_cipher *tfm;
        tfm = crypto_alloc_cipher("aes", 0, CRYPTO_ALG_ASYNC);
-        if (IS_ERR(tfm))
+        if (!IS_ERR(tfm))
-                return NULL;
+                crypto_cipher_setkey(tfm, key, ALG_CCMP_KEY_LEN);
-        crypto_cipher_setkey(tfm, key, ALG_CCMP_KEY_LEN);
        return tfm;
 }
diff --git a/net/mac80211/aes_cmac.c b/net/mac80211/aes_cmac.c
index 3d097b3d7b62..b4d66cca76d6 100644
--- a/net/mac80211/aes_cmac.c
+++ b/net/mac80211/aes_cmac.c
@@ -119,10 +119,8 @@ struct crypto_cipher * ieee80211_aes_cmac_key_setup(const u8 key[])
        struct crypto_cipher *tfm;
        tfm = crypto_alloc_cipher("aes", 0, CRYPTO_ALG_ASYNC);
-        if (IS_ERR(tfm))
+        if (!IS_ERR(tfm))
-                return NULL;
+                crypto_cipher_setkey(tfm, key, AES_CMAC_KEY_LEN);
-        crypto_cipher_setkey(tfm, key, AES_CMAC_KEY_LEN);
        return tfm;
 }
diff --git a/net/mac80211/agg-rx.c b/net/mac80211/agg-rx.c
index 965b272499fd..58eab9e8e4ee 100644
--- a/net/mac80211/agg-rx.c
+++ b/net/mac80211/agg-rx.c
@@ -86,6 +86,7 @@ void ___ieee80211_stop_rx_ba_session(struct sta_info *sta, u16 tid,
                                     tid, 0, reason);
        del_timer_sync(&tid_rx->session_timer);
+        del_timer_sync(&tid_rx->reorder_timer);
        call_rcu(&tid_rx->rcu_head, ieee80211_free_tid_rx);
 }
@@ -120,6 +121,20 @@ static void sta_rx_agg_session_timer_expired(unsigned long data)
        ieee80211_queue_work(&sta->local->hw, &sta->ampdu_mlme.work);
 }
+static void sta_rx_agg_reorder_timer_expired(unsigned long data)
+{
+        u8 *ptid = (u8 *)data;
+        u8 *timer_to_id = ptid - *ptid;
+        struct sta_info *sta = container_of(timer_to_id, struct sta_info,
+                        timer_to_tid[0]);
+        rcu_read_lock();
+        spin_lock(&sta->lock);
+        ieee80211_release_reorder_timeout(sta, *ptid);
+        spin_unlock(&sta->lock);
+        rcu_read_unlock();
+}
 static void ieee80211_send_addba_resp(struct ieee80211_sub_if_data *sdata, u8 *da, u16 tid,
                                      u8 dialog_token, u16 status, u16 policy,
                                      u16 buf_size, u16 timeout)
@@ -251,11 +266,18 @@ void ieee80211_process_addba_request(struct ieee80211_local *local,
                goto end;
        }
+        spin_lock_init(&tid_agg_rx->reorder_lock);
        /* rx timer */
        tid_agg_rx->session_timer.function = sta_rx_agg_session_timer_expired;
        tid_agg_rx->session_timer.data = (unsigned long)&sta->timer_to_tid[tid];
        init_timer(&tid_agg_rx->session_timer);
+        /* rx reorder timer */
+        tid_agg_rx->reorder_timer.function = sta_rx_agg_reorder_timer_expired;
+        tid_agg_rx->reorder_timer.data = (unsigned long)&sta->timer_to_tid[tid];
+        init_timer(&tid_agg_rx->reorder_timer);
        /* prepare reordering buffer */
        tid_agg_rx->reorder_buf =
                kcalloc(buf_size, sizeof(struct sk_buff *), GFP_ATOMIC);
diff --git a/net/mac80211/cfg.c b/net/mac80211/cfg.c
index 29ac8e1a509e..c981604b71e6 100644
--- a/net/mac80211/cfg.c
+++ b/net/mac80211/cfg.c
@@ -19,33 +19,6 @@
 #include "rate.h"
 #include "mesh.h"
-static bool nl80211_type_check(enum nl80211_iftype type)
-{
-        switch (type) {
-        case NL80211_IFTYPE_ADHOC:
-        case NL80211_IFTYPE_STATION:
-        case NL80211_IFTYPE_MONITOR:
-#ifdef CONFIG_MAC80211_MESH
-        case NL80211_IFTYPE_MESH_POINT:
-#endif
-        case NL80211_IFTYPE_AP:
-        case NL80211_IFTYPE_AP_VLAN:
-        case NL80211_IFTYPE_WDS:
-                return true;
-        default:
-                return false;
-        }
-}
-static bool nl80211_params_check(enum nl80211_iftype type,
-                                 struct vif_params *params)
-{
-        if (!nl80211_type_check(type))
-                return false;
-        return true;
-}
 static int ieee80211_add_iface(struct wiphy *wiphy, char *name,
                               enum nl80211_iftype type, u32 *flags,
                               struct vif_params *params)
@@ -55,9 +28,6 @@ static int ieee80211_add_iface(struct wiphy *wiphy, char *name,
        struct ieee80211_sub_if_data *sdata;
        int err;
-        if (!nl80211_params_check(type, params))
-                return -EINVAL;
        err = ieee80211_if_add(local, name, &dev, type, params);
        if (err || type != NL80211_IFTYPE_MONITOR || !flags)
                return err;
@@ -82,12 +52,6 @@ static int ieee80211_change_iface(struct wiphy *wiphy,
        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
        int ret;
-        if (ieee80211_sdata_running(sdata))
-                return -EBUSY;
-        if (!nl80211_params_check(type, params))
-                return -EINVAL;
        ret = ieee80211_if_change_type(sdata, type);
        if (ret)
                return ret;
@@ -114,44 +78,30 @@ static int ieee80211_add_key(struct wiphy *wiphy, struct net_device *dev,
                             u8 key_idx, const u8 *mac_addr,
                             struct key_params *params)
 {
-        struct ieee80211_sub_if_data *sdata;
+        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
        struct sta_info *sta = NULL;
-        enum ieee80211_key_alg alg;
        struct ieee80211_key *key;
        int err;
-        if (!netif_running(dev))
+        if (!ieee80211_sdata_running(sdata))
                return -ENETDOWN;
-        sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+        /* reject WEP and TKIP keys if WEP failed to initialize */
        switch (params->cipher) {
        case WLAN_CIPHER_SUITE_WEP40:
-        case WLAN_CIPHER_SUITE_WEP104:
-                alg = ALG_WEP;
-                break;
        case WLAN_CIPHER_SUITE_TKIP:
-                alg = ALG_TKIP;
+        case WLAN_CIPHER_SUITE_WEP104:
-                break;
+                if (IS_ERR(sdata->local->wep_tx_tfm))
-        case WLAN_CIPHER_SUITE_CCMP:
+                        return -EINVAL;
-                alg = ALG_CCMP;
-                break;
-        case WLAN_CIPHER_SUITE_AES_CMAC:
-                alg = ALG_AES_CMAC;
                break;
        default:
-                return -EINVAL;
+                break;
        }
-        /* reject WEP and TKIP keys if WEP failed to initialize */
+        key = ieee80211_key_alloc(params->cipher, key_idx, params->key_len,
-        if ((alg == ALG_WEP || alg == ALG_TKIP) &&
+                                  params->key, params->seq_len, params->seq);
-            IS_ERR(sdata->local->wep_tx_tfm))
+        if (IS_ERR(key))
-                return -EINVAL;
+                return PTR_ERR(key);
-        key = ieee80211_key_alloc(alg, key_idx, params->key_len, params->key,
-                                  params->seq_len, params->seq);
-        if (!key)
-                return -ENOMEM;
        mutex_lock(&sdata->local->sta_mtx);
@@ -164,9 +114,10 @@ static int ieee80211_add_key(struct wiphy *wiphy, struct net_device *dev,
                }
        }
-        ieee80211_key_link(key, sdata, sta);
+        err = ieee80211_key_link(key, sdata, sta);
+        if (err)
+                ieee80211_key_free(sdata->local, key);
-        err = 0;
 out_unlock:
        mutex_unlock(&sdata->local->sta_mtx);
@@ -247,10 +198,10 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev,
        memset(&params, 0, sizeof(params));
-        switch (key->conf.alg) {
+        params.cipher = key->conf.cipher;
-        case ALG_TKIP:
-                params.cipher = WLAN_CIPHER_SUITE_TKIP;
+        switch (key->conf.cipher) {
+        case WLAN_CIPHER_SUITE_TKIP:
                iv32 = key->u.tkip.tx.iv32;
                iv16 = key->u.tkip.tx.iv16;
@@ -268,8 +219,7 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev,
                params.seq = seq;
                params.seq_len = 6;
                break;
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
-                params.cipher = WLAN_CIPHER_SUITE_CCMP;
                seq[0] = key->u.ccmp.tx_pn[5];
                seq[1] = key->u.ccmp.tx_pn[4];
                seq[2] = key->u.ccmp.tx_pn[3];
@@ -279,14 +229,7 @@ static int ieee80211_get_key(struct wiphy *wiphy, struct net_device *dev,
                params.seq = seq;
                params.seq_len = 6;
                break;
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
-                if (key->conf.keylen == 5)
-                        params.cipher = WLAN_CIPHER_SUITE_WEP40;
-                else
-                        params.cipher = WLAN_CIPHER_SUITE_WEP104;
-                break;
-        case ALG_AES_CMAC:
-                params.cipher = WLAN_CIPHER_SUITE_AES_CMAC;
                seq[0] = key->u.aes_cmac.tx_pn[5];
                seq[1] = key->u.aes_cmac.tx_pn[4];
                seq[2] = key->u.aes_cmac.tx_pn[3];
@@ -634,6 +577,7 @@ static void sta_apply_parameters(struct ieee80211_local *local,
                                 struct sta_info *sta,
                                 struct station_parameters *params)
 {
+        unsigned long flags;
        u32 rates;
        int i, j;
        struct ieee80211_supported_band *sband;
@@ -642,7 +586,7 @@ static void sta_apply_parameters(struct ieee80211_local *local,
        sband = local->hw.wiphy->bands[local->oper_channel->band];
-        spin_lock_bh(&sta->lock);
+        spin_lock_irqsave(&sta->flaglock, flags);
        mask = params->sta_flags_mask;
        set = params->sta_flags_set;
@@ -669,7 +613,7 @@ static void sta_apply_parameters(struct ieee80211_local *local,
                if (set & BIT(NL80211_STA_FLAG_MFP))
                        sta->flags |= WLAN_STA_MFP;
        }
-        spin_unlock_bh(&sta->lock);
+        spin_unlock_irqrestore(&sta->flaglock, flags);
        /*
         * cfg80211 validates this (1-2007) and allows setting the AID
@@ -1143,9 +1087,9 @@ static int ieee80211_set_txq_params(struct wiphy *wiphy,
        p.uapsd = false;
        if (drv_conf_tx(local, params->queue, &p)) {
-                printk(KERN_DEBUG "%s: failed to set TX queue "
+                wiphy_debug(local->hw.wiphy,
-                       "parameters for queue %d\n",
+                            "failed to set TX queue parameters for queue %d\n",
-                       wiphy_name(local->hw.wiphy), params->queue);
+                            params->queue);
                return -EINVAL;
        }
@@ -1207,15 +1151,26 @@ static int ieee80211_scan(struct wiphy *wiphy,
                          struct net_device *dev,
                          struct cfg80211_scan_request *req)
 {
-        struct ieee80211_sub_if_data *sdata;
+        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
-        sdata = IEEE80211_DEV_TO_SUB_IF(dev);
-        if (sdata->vif.type != NL80211_IFTYPE_STATION &&
+        switch (ieee80211_vif_type_p2p(&sdata->vif)) {
-            sdata->vif.type != NL80211_IFTYPE_ADHOC &&
+        case NL80211_IFTYPE_STATION:
-            sdata->vif.type != NL80211_IFTYPE_MESH_POINT &&
+        case NL80211_IFTYPE_ADHOC:
-            (sdata->vif.type != NL80211_IFTYPE_AP || sdata->u.ap.beacon))
+        case NL80211_IFTYPE_MESH_POINT:
+        case NL80211_IFTYPE_P2P_CLIENT:
+                break;
+        case NL80211_IFTYPE_P2P_GO:
+                if (sdata->local->ops->hw_scan)
+                        break;
+                /* FIXME: implement NoA while scanning in software */
+                return -EOPNOTSUPP;
+        case NL80211_IFTYPE_AP:
+                if (sdata->u.ap.beacon)
+                        return -EOPNOTSUPP;
+                break;
+        default:
                return -EOPNOTSUPP;
+        }
        return ieee80211_request_scan(sdata, req);
 }
@@ -1541,11 +1496,11 @@ static int ieee80211_cancel_remain_on_channel(struct wiphy *wiphy,
        return ieee80211_wk_cancel_remain_on_channel(sdata, cookie);
 }
-static int ieee80211_action(struct wiphy *wiphy, struct net_device *dev,
+static int ieee80211_mgmt_tx(struct wiphy *wiphy, struct net_device *dev,
-                            struct ieee80211_channel *chan,
+                             struct ieee80211_channel *chan,
-                            enum nl80211_channel_type channel_type,
+                             enum nl80211_channel_type channel_type,
-                            bool channel_type_valid,
+                             bool channel_type_valid,
-                            const u8 *buf, size_t len, u64 *cookie)
+                             const u8 *buf, size_t len, u64 *cookie)
 {
        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
        struct ieee80211_local *local = sdata->local;
@@ -1575,8 +1530,6 @@ static int ieee80211_action(struct wiphy *wiphy, struct net_device *dev,
                        return -ENOLINK;
                break;
        case NL80211_IFTYPE_STATION:
-                if (!(sdata->u.mgd.flags & IEEE80211_STA_MFP_ENABLED))
-                        flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
                break;
        default:
                return -EOPNOTSUPP;
@@ -1647,6 +1600,6 @@ struct cfg80211_ops mac80211_config_ops = {
        .set_bitrate_mask = ieee80211_set_bitrate_mask,
        .remain_on_channel = ieee80211_remain_on_channel,
        .cancel_remain_on_channel = ieee80211_cancel_remain_on_channel,
-        .action = ieee80211_action,
+        .mgmt_tx = ieee80211_mgmt_tx,
        .set_cqm_rssi_config = ieee80211_set_cqm_rssi_config,
 };
diff --git a/net/mac80211/chan.c b/net/mac80211/chan.c
index 32be11e4c4d9..5b24740fc0b0 100644
--- a/net/mac80211/chan.c
+++ b/net/mac80211/chan.c
@@ -11,7 +11,7 @@ __ieee80211_get_channel_mode(struct ieee80211_local *local,
 {
        struct ieee80211_sub_if_data *sdata;
-        WARN_ON(!mutex_is_locked(&local->iflist_mtx));
+        lockdep_assert_held(&local->iflist_mtx);
        list_for_each_entry(sdata, &local->interfaces, list) {
                if (sdata == ignore)
diff --git a/net/mac80211/debugfs.c b/net/mac80211/debugfs.c
index a694c593ff6a..ebd5b69f562e 100644
--- a/net/mac80211/debugfs.c
+++ b/net/mac80211/debugfs.c
@@ -85,13 +85,15 @@ static ssize_t tsf_write(struct file *file,
        if (strncmp(buf, "reset", 5) == 0) {
                if (local->ops->reset_tsf) {
                        drv_reset_tsf(local);
-                        printk(KERN_INFO "%s: debugfs reset TSF\n", wiphy_name(local->hw.wiphy));
+                        wiphy_info(local->hw.wiphy, "debugfs reset TSF\n");
                }
        } else {
                tsf = simple_strtoul(buf, NULL, 0);
                if (local->ops->set_tsf) {
                        drv_set_tsf(local, tsf);
-                        printk(KERN_INFO "%s: debugfs set TSF to %#018llx\n", wiphy_name(local->hw.wiphy), tsf);
+                        wiphy_info(local->hw.wiphy,
+                                   "debugfs set TSF to %#018llx\n", tsf);
                }
        }
@@ -366,7 +368,6 @@ void debugfs_hw_add(struct ieee80211_local *local)
        if (!phyd)
                return;
-        local->debugfs.stations = debugfs_create_dir("stations", phyd);
        local->debugfs.keys = debugfs_create_dir("keys", phyd);
        DEBUGFS_ADD(frequency);
diff --git a/net/mac80211/debugfs_key.c b/net/mac80211/debugfs_key.c
index fa5e76e658ef..1647f8dc5cda 100644
--- a/net/mac80211/debugfs_key.c
+++ b/net/mac80211/debugfs_key.c
@@ -64,26 +64,13 @@ static ssize_t key_algorithm_read(struct file *file,
                                  char __user *userbuf,
                                  size_t count, loff_t *ppos)
 {
-        char *alg;
+        char buf[15];
        struct ieee80211_key *key = file->private_data;
+        u32 c = key->conf.cipher;
-        switch (key->conf.alg) {
+        sprintf(buf, "%.2x-%.2x-%.2x:%d\n",
-        case ALG_WEP:
+                c >> 24, (c >> 16) & 0xff, (c >> 8) & 0xff, c & 0xff);
-                alg = "WEP\n";
+        return simple_read_from_buffer(userbuf, count, ppos, buf, strlen(buf));
-                break;
-        case ALG_TKIP:
-                alg = "TKIP\n";
-                break;
-        case ALG_CCMP:
-                alg = "CCMP\n";
-                break;
-        case ALG_AES_CMAC:
-                alg = "AES-128-CMAC\n";
-                break;
-        default:
-                return 0;
-        }
-        return simple_read_from_buffer(userbuf, count, ppos, alg, strlen(alg));
 }
 KEY_OPS(algorithm);
@@ -95,21 +82,22 @@ static ssize_t key_tx_spec_read(struct file *file, char __user *userbuf,
        int len;
        struct ieee80211_key *key = file->private_data;
-        switch (key->conf.alg) {
+        switch (key->conf.cipher) {
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_WEP40:
+        case WLAN_CIPHER_SUITE_WEP104:
                len = scnprintf(buf, sizeof(buf), "\n");
                break;
-        case ALG_TKIP:
+        case WLAN_CIPHER_SUITE_TKIP:
                len = scnprintf(buf, sizeof(buf), "%08x %04x\n",
                                key->u.tkip.tx.iv32,
                                key->u.tkip.tx.iv16);
                break;
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                tpn = key->u.ccmp.tx_pn;
                len = scnprintf(buf, sizeof(buf), "%02x%02x%02x%02x%02x%02x\n",
                                tpn[0], tpn[1], tpn[2], tpn[3], tpn[4], tpn[5]);
                break;
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                tpn = key->u.aes_cmac.tx_pn;
                len = scnprintf(buf, sizeof(buf), "%02x%02x%02x%02x%02x%02x\n",
                                tpn[0], tpn[1], tpn[2], tpn[3], tpn[4],
@@ -130,11 +118,12 @@ static ssize_t key_rx_spec_read(struct file *file, char __user *userbuf,
        int i, len;
        const u8 *rpn;
-        switch (key->conf.alg) {
+        switch (key->conf.cipher) {
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_WEP40:
+        case WLAN_CIPHER_SUITE_WEP104:
                len = scnprintf(buf, sizeof(buf), "\n");
                break;
-        case ALG_TKIP:
+        case WLAN_CIPHER_SUITE_TKIP:
                for (i = 0; i < NUM_RX_DATA_QUEUES; i++)
                        p += scnprintf(p, sizeof(buf)+buf-p,
                                       "%08x %04x\n",
@@ -142,7 +131,7 @@ static ssize_t key_rx_spec_read(struct file *file, char __user *userbuf,
                                       key->u.tkip.rx[i].iv16);
                len = p - buf;
                break;
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                for (i = 0; i < NUM_RX_DATA_QUEUES + 1; i++) {
                        rpn = key->u.ccmp.rx_pn[i];
                        p += scnprintf(p, sizeof(buf)+buf-p,
@@ -152,7 +141,7 @@ static ssize_t key_rx_spec_read(struct file *file, char __user *userbuf,
                }
                len = p - buf;
                break;
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                rpn = key->u.aes_cmac.rx_pn;
                p += scnprintf(p, sizeof(buf)+buf-p,
                               "%02x%02x%02x%02x%02x%02x\n",
@@ -174,11 +163,11 @@ static ssize_t key_replays_read(struct file *file, char __user *userbuf,
        char buf[20];
        int len;
-        switch (key->conf.alg) {
+        switch (key->conf.cipher) {
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                len = scnprintf(buf, sizeof(buf), "%u\n", key->u.ccmp.replays);
                break;
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                len = scnprintf(buf, sizeof(buf), "%u\n",
                                key->u.aes_cmac.replays);
                break;
@@ -196,8 +185,8 @@ static ssize_t key_icverrors_read(struct file *file, char __user *userbuf,
        char buf[20];
        int len;
-        switch (key->conf.alg) {
+        switch (key->conf.cipher) {
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                len = scnprintf(buf, sizeof(buf), "%u\n",
                                key->u.aes_cmac.icverrors);
                break;
diff --git a/net/mac80211/debugfs_netdev.c b/net/mac80211/debugfs_netdev.c
index 20b2998fa0ed..3e12430591b7 100644
--- a/net/mac80211/debugfs_netdev.c
+++ b/net/mac80211/debugfs_netdev.c
@@ -409,6 +409,9 @@ void ieee80211_debugfs_add_netdev(struct ieee80211_sub_if_data *sdata)
        sprintf(buf, "netdev:%s", sdata->name);
        sdata->debugfs.dir = debugfs_create_dir(buf,
                sdata->local->hw.wiphy->debugfsdir);
+        if (sdata->debugfs.dir)
+                sdata->debugfs.subdir_stations = debugfs_create_dir("stations",
+                        sdata->debugfs.dir);
        add_files(sdata);
 }
diff --git a/net/mac80211/debugfs_sta.c b/net/mac80211/debugfs_sta.c
index 76839d4dfaac..6b7ff9fb4604 100644
--- a/net/mac80211/debugfs_sta.c
+++ b/net/mac80211/debugfs_sta.c
@@ -300,7 +300,7 @@ STA_OPS(ht_capa);
 void ieee80211_sta_debugfs_add(struct sta_info *sta)
 {
-        struct dentry *stations_dir = sta->local->debugfs.stations;
+        struct dentry *stations_dir = sta->sdata->debugfs.subdir_stations;
        u8 mac[3*ETH_ALEN];
        sta->debugfs.add_has_run = true;
diff --git a/net/mac80211/driver-ops.h b/net/mac80211/driver-ops.h
index 14123dce544b..16983825f8e8 100644
--- a/net/mac80211/driver-ops.h
+++ b/net/mac80211/driver-ops.h
@@ -54,6 +54,20 @@ static inline int drv_add_interface(struct ieee80211_local *local,
        return ret;
 }
+static inline int drv_change_interface(struct ieee80211_local *local,
+                                       struct ieee80211_sub_if_data *sdata,
+                                       enum nl80211_iftype type, bool p2p)
+{
+        int ret;
+        might_sleep();
+        trace_drv_change_interface(local, sdata, type, p2p);
+        ret = local->ops->change_interface(&local->hw, &sdata->vif, type, p2p);
+        trace_drv_return_int(local, ret);
+        return ret;
+}
 static inline void drv_remove_interface(struct ieee80211_local *local,
                                        struct ieee80211_vif *vif)
 {
diff --git a/net/mac80211/driver-trace.h b/net/mac80211/driver-trace.h
index 5d5d2a974668..6831fb1641c8 100644
--- a/net/mac80211/driver-trace.h
+++ b/net/mac80211/driver-trace.h
@@ -25,12 +25,14 @@ static inline void trace_ ## name(proto) {}
 #define STA_PR_FMT      " sta:%pM"
 #define STA_PR_ARG      __entry->sta_addr
-#define VIF_ENTRY       __field(enum nl80211_iftype, vif_type) __field(void *, sdata) \
+#define VIF_ENTRY       __field(enum nl80211_iftype, vif_type) __field(void *, sdata)   \
+                        __field(bool, p2p)                                              \
                        __string(vif_name, sdata->dev ? sdata->dev->name : "<nodev>")
-#define VIF_ASSIGN      __entry->vif_type = sdata->vif.type; __entry->sdata = sdata; \
+#define VIF_ASSIGN      __entry->vif_type = sdata->vif.type; __entry->sdata = sdata;    \
+                        __entry->p2p = sdata->vif.p2p;                                  \
                        __assign_str(vif_name, sdata->dev ? sdata->dev->name : "<nodev>")
-#define VIF_PR_FMT      " vif:%s(%d)"
+#define VIF_PR_FMT      " vif:%s(%d%s)"
-#define VIF_PR_ARG      __get_str(vif_name), __entry->vif_type
+#define VIF_PR_ARG      __get_str(vif_name), __entry->vif_type, __entry->p2p ? "/p2p" : ""
 /*
 * Tracing for driver callbacks.
@@ -136,6 +138,34 @@ TRACE_EVENT(drv_add_interface,
        )
 );
+TRACE_EVENT(drv_change_interface,
+        TP_PROTO(struct ieee80211_local *local,
+                 struct ieee80211_sub_if_data *sdata,
+                 enum nl80211_iftype type, bool p2p),
+        TP_ARGS(local, sdata, type, p2p),
+        TP_STRUCT__entry(
+                LOCAL_ENTRY
+                VIF_ENTRY
+                __field(u32, new_type)
+                __field(bool, new_p2p)
+        ),
+        TP_fast_assign(
+                LOCAL_ASSIGN;
+                VIF_ASSIGN;
+                __entry->new_type = type;
+                __entry->new_p2p = p2p;
+        ),
+        TP_printk(
+                LOCAL_PR_FMT  VIF_PR_FMT " new type:%d%s",
+                LOCAL_PR_ARG, VIF_PR_ARG, __entry->new_type,
+                __entry->new_p2p ? "/p2p" : ""
+        )
+);
 TRACE_EVENT(drv_remove_interface,
        TP_PROTO(struct ieee80211_local *local, struct ieee80211_sub_if_data *sdata),
@@ -336,7 +366,7 @@ TRACE_EVENT(drv_set_key,
                LOCAL_ENTRY
                VIF_ENTRY
                STA_ENTRY
-                __field(enum ieee80211_key_alg, alg)
+                __field(u32, cipher)
                __field(u8, hw_key_idx)
                __field(u8, flags)
                __field(s8, keyidx)
@@ -346,7 +376,7 @@ TRACE_EVENT(drv_set_key,
                LOCAL_ASSIGN;
                VIF_ASSIGN;
                STA_ASSIGN;
-                __entry->alg = key->alg;
+                __entry->cipher = key->cipher;
                __entry->flags = key->flags;
                __entry->keyidx = key->keyidx;
                __entry->hw_key_idx = key->hw_key_idx;
diff --git a/net/mac80211/ht.c b/net/mac80211/ht.c
index 9d101fb33861..11f74f5f7b2f 100644
--- a/net/mac80211/ht.c
+++ b/net/mac80211/ht.c
@@ -265,3 +265,31 @@ int ieee80211_send_smps_action(struct ieee80211_sub_if_data *sdata,
        return 0;
 }
+void ieee80211_request_smps_work(struct work_struct *work)
+{
+        struct ieee80211_sub_if_data *sdata =
+                container_of(work, struct ieee80211_sub_if_data,
+                             u.mgd.request_smps_work);
+        mutex_lock(&sdata->u.mgd.mtx);
+        __ieee80211_request_smps(sdata, sdata->u.mgd.driver_smps_mode);
+        mutex_unlock(&sdata->u.mgd.mtx);
+}
+void ieee80211_request_smps(struct ieee80211_vif *vif,
+                            enum ieee80211_smps_mode smps_mode)
+{
+        struct ieee80211_sub_if_data *sdata = vif_to_sdata(vif);
+        if (WARN_ON(vif->type != NL80211_IFTYPE_STATION))
+                return;
+        if (WARN_ON(smps_mode == IEEE80211_SMPS_OFF))
+                smps_mode = IEEE80211_SMPS_AUTOMATIC;
+        ieee80211_queue_work(&sdata->local->hw,
+                             &sdata->u.mgd.request_smps_work);
+}
+/* this might change ... don't want non-open drivers using it */
+EXPORT_SYMBOL_GPL(ieee80211_request_smps);
diff --git a/net/mac80211/ibss.c b/net/mac80211/ibss.c
index c691780725a7..1a3aae54f0cf 100644
--- a/net/mac80211/ibss.c
+++ b/net/mac80211/ibss.c
@@ -427,8 +427,8 @@ struct sta_info *ieee80211_ibss_add_sta(struct ieee80211_sub_if_data *sdata,
                return NULL;
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: Adding new IBSS station %pM (dev=%s)\n",
+        wiphy_debug(local->hw.wiphy, "Adding new IBSS station %pM (dev=%s)\n",
-               wiphy_name(local->hw.wiphy), addr, sdata->name);
+                    addr, sdata->name);
 #endif
        sta = sta_info_alloc(sdata, addr, gfp);
@@ -920,12 +920,14 @@ int ieee80211_ibss_join(struct ieee80211_sub_if_data *sdata,
        memcpy(sdata->u.ibss.ssid, params->ssid, IEEE80211_MAX_SSID_LEN);
        sdata->u.ibss.ssid_len = params->ssid_len;
+        mutex_unlock(&sdata->u.ibss.mtx);
+        mutex_lock(&sdata->local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&sdata->local->mtx);
        ieee80211_queue_work(&sdata->local->hw, &sdata->work);
-        mutex_unlock(&sdata->u.ibss.mtx);
        return 0;
 }
@@ -980,7 +982,9 @@ int ieee80211_ibss_leave(struct ieee80211_sub_if_data *sdata)
        mutex_unlock(&sdata->u.ibss.mtx);
+        mutex_lock(&local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&local->mtx);
        return 0;
 }
diff --git a/net/mac80211/ieee80211_i.h b/net/mac80211/ieee80211_i.h
index 65e0ed6c2975..945fbf29719d 100644
--- a/net/mac80211/ieee80211_i.h
+++ b/net/mac80211/ieee80211_i.h
@@ -50,12 +50,6 @@ struct ieee80211_local;
 * increased memory use (about 2 kB of RAM per entry). */
 #define IEEE80211_FRAGMENT_MAX 4
-/*
- * Time after which we ignore scan results and no longer report/use
- * them in any way.
- */
-#define IEEE80211_SCAN_RESULT_EXPIRE (10 * HZ)
 #define TU_TO_EXP_TIME(x)       (jiffies + usecs_to_jiffies((x) * 1024))
 #define IEEE80211_DEFAULT_UAPSD_QUEUES \
@@ -165,12 +159,37 @@ typedef unsigned __bitwise__ ieee80211_rx_result;
 #define RX_DROP_MONITOR         ((__force ieee80211_rx_result) 2u)
 #define RX_QUEUED               ((__force ieee80211_rx_result) 3u)
-#define IEEE80211_RX_IN_SCAN            BIT(0)
+/**
-/* frame is destined to interface currently processed (incl. multicast frames) */
+ * enum ieee80211_packet_rx_flags - packet RX flags
-#define IEEE80211_RX_RA_MATCH           BIT(1)
+ * @IEEE80211_RX_RA_MATCH: frame is destined to interface currently processed
-#define IEEE80211_RX_AMSDU              BIT(2)
+ *      (incl. multicast frames)
-#define IEEE80211_RX_FRAGMENTED         BIT(3)
+ * @IEEE80211_RX_IN_SCAN: received while scanning
-/* only add flags here that do not change with subframes of an aMPDU */
+ * @IEEE80211_RX_FRAGMENTED: fragmented frame
+ * @IEEE80211_RX_AMSDU: a-MSDU packet
+ * @IEEE80211_RX_MALFORMED_ACTION_FRM: action frame is malformed
+ *
+ * These are per-frame flags that are attached to a frame in the
+ * @rx_flags field of &struct ieee80211_rx_status.
+ */
+enum ieee80211_packet_rx_flags {
+        IEEE80211_RX_IN_SCAN                    = BIT(0),
+        IEEE80211_RX_RA_MATCH                   = BIT(1),
+        IEEE80211_RX_FRAGMENTED                 = BIT(2),
+        IEEE80211_RX_AMSDU                      = BIT(3),
+        IEEE80211_RX_MALFORMED_ACTION_FRM       = BIT(4),
+};
+/**
+ * enum ieee80211_rx_flags - RX data flags
+ *
+ * @IEEE80211_RX_CMNTR: received on cooked monitor already
+ *
+ * These flags are used across handling multiple interfaces
+ * for a single frame.
+ */
+enum ieee80211_rx_flags {
+        IEEE80211_RX_CMNTR              = BIT(0),
+};
 struct ieee80211_rx_data {
        struct sk_buff *skb;
@@ -343,7 +362,10 @@ struct ieee80211_if_managed {
        unsigned long timers_running; /* used for quiesce/restart */
        bool powersave; /* powersave requested for this iface */
        enum ieee80211_smps_mode req_smps, /* requested smps mode */
-                                 ap_smps; /* smps mode AP thinks we're in */
+                                 ap_smps, /* smps mode AP thinks we're in */
+                                 driver_smps_mode; /* smps mode request */
+        struct work_struct request_smps_work;
        unsigned int flags;
@@ -371,6 +393,13 @@ struct ieee80211_if_managed {
        int ave_beacon_signal;
        /*
+         * Number of Beacon frames used in ave_beacon_signal. This can be used
+         * to avoid generating less reliable cqm events that would be based
+         * only on couple of received frames.
+         */
+        unsigned int count_beacon_signal;
+        /*
         * Last Beacon frame signal strength average (ave_beacon_signal / 16)
         * that triggered a cqm event. 0 indicates that no event has been
         * generated for the current association.
@@ -474,6 +503,19 @@ enum ieee80211_sub_if_data_flags {
        IEEE80211_SDATA_DONT_BRIDGE_PACKETS     = BIT(3),
 };
+/**
+ * enum ieee80211_sdata_state_bits - virtual interface state bits
+ * @SDATA_STATE_RUNNING: virtual interface is up & running; this
+ *      mirrors netif_running() but is separate for interface type
+ *      change handling while the interface is up
+ * @SDATA_STATE_OFFCHANNEL: This interface is currently in offchannel
+ *      mode, so queues are stopped
+ */
+enum ieee80211_sdata_state_bits {
+        SDATA_STATE_RUNNING,
+        SDATA_STATE_OFFCHANNEL,
+};
 struct ieee80211_sub_if_data {
        struct list_head list;
@@ -487,6 +529,8 @@ struct ieee80211_sub_if_data {
        unsigned int flags;
+        unsigned long state;
        int drop_unencrypted;
        char name[IFNAMSIZ];
@@ -497,6 +541,9 @@ struct ieee80211_sub_if_data {
         */
        bool ht_opmode_valid;
+        /* to detect idle changes */
+        bool old_idle;
        /* Fragment table for host-based reassembly */
        struct ieee80211_fragment_entry fragments[IEEE80211_FRAGMENT_MAX];
        unsigned int fragment_next;
@@ -508,6 +555,8 @@ struct ieee80211_sub_if_data {
        struct ieee80211_key *default_mgmt_key;
        u16 sequence_number;
+        __be16 control_port_protocol;
+        bool control_port_no_encrypt;
        struct work_struct work;
        struct sk_buff_head skb_queue;
@@ -539,6 +588,7 @@ struct ieee80211_sub_if_data {
 #ifdef CONFIG_MAC80211_DEBUGFS
        struct {
                struct dentry *dir;
+                struct dentry *subdir_stations;
                struct dentry *default_key;
                struct dentry *default_mgmt_key;
        } debugfs;
@@ -595,11 +645,17 @@ enum queue_stop_reason {
 *      determine if we are on the operating channel or not
 * @SCAN_OFF_CHANNEL: We're off our operating channel for scanning,
 *      gets only set in conjunction with SCAN_SW_SCANNING
+ * @SCAN_COMPLETED: Set for our scan work function when the driver reported
+ *      that the scan completed.
+ * @SCAN_ABORTED: Set for our scan work function when the driver reported
+ *      a scan complete for an aborted scan.
 */
 enum {
        SCAN_SW_SCANNING,
        SCAN_HW_SCANNING,
        SCAN_OFF_CHANNEL,
+        SCAN_COMPLETED,
+        SCAN_ABORTED,
 };
 /**
@@ -634,7 +690,6 @@ struct ieee80211_local {
        /*
         * work stuff, potentially off-channel (in the future)
         */
-        struct mutex work_mtx;
        struct list_head work_list;
        struct timer_list work_timer;
        struct work_struct work_work;
@@ -656,6 +711,8 @@ struct ieee80211_local {
        int fif_fcsfail, fif_plcpfail, fif_control, fif_other_bss, fif_pspoll;
        unsigned int filter_flags; /* FIF_* */
+        bool wiphy_ciphers_allocated;
        /* protects the aggregated multicast list and filter calls */
        spinlock_t filter_lock;
@@ -746,9 +803,10 @@ struct ieee80211_local {
         */
        struct mutex key_mtx;
+        /* mutex for scan and work locking */
+        struct mutex mtx;
        /* Scanning and BSS list */
-        struct mutex scan_mtx;
        unsigned long scanning;
        struct cfg80211_ssid scan_ssid;
        struct cfg80211_scan_request *int_scan_req;
@@ -866,10 +924,14 @@ struct ieee80211_local {
 #ifdef CONFIG_MAC80211_DEBUGFS
        struct local_debugfsdentries {
                struct dentry *rcdir;
-                struct dentry *stations;
                struct dentry *keys;
        } debugfs;
 #endif
+        /* dummy netdev for use w/ NAPI */
+        struct net_device napi_dev;
+        struct napi_struct napi;
 };
 static inline struct ieee80211_sub_if_data *
@@ -1003,6 +1065,8 @@ void ieee80211_sta_restart(struct ieee80211_sub_if_data *sdata);
 void ieee80211_sta_work(struct ieee80211_sub_if_data *sdata);
 void ieee80211_sta_rx_queued_mgmt(struct ieee80211_sub_if_data *sdata,
                                  struct sk_buff *skb);
+void ieee80211_sta_reset_beacon_monitor(struct ieee80211_sub_if_data *sdata);
+void ieee80211_sta_reset_conn_monitor(struct ieee80211_sub_if_data *sdata);
 /* IBSS code */
 void ieee80211_ibss_notify_scan_completed(struct ieee80211_local *local);
@@ -1071,7 +1135,7 @@ void ieee80211_recalc_idle(struct ieee80211_local *local);
 static inline bool ieee80211_sdata_running(struct ieee80211_sub_if_data *sdata)
 {
-        return netif_running(sdata->dev);
+        return test_bit(SDATA_STATE_RUNNING, &sdata->state);
 }
 /* tx handling */
@@ -1105,6 +1169,7 @@ void ieee80211_send_delba(struct ieee80211_sub_if_data *sdata,
 int ieee80211_send_smps_action(struct ieee80211_sub_if_data *sdata,
                               enum ieee80211_smps_mode smps, const u8 *da,
                               const u8 *bssid);
+void ieee80211_request_smps_work(struct work_struct *work);
 void ___ieee80211_stop_rx_ba_session(struct sta_info *sta, u16 tid,
                                     u16 initiator, u16 reason);
@@ -1131,6 +1196,7 @@ void ieee80211_start_tx_ba_cb(struct ieee80211_vif *vif, u8 *ra, u16 tid);
 void ieee80211_stop_tx_ba_cb(struct ieee80211_vif *vif, u8 *ra, u8 tid);
 void ieee80211_ba_session_work(struct work_struct *work);
 void ieee80211_tx_ba_session_handle_start(struct sta_info *sta, int tid);
+void ieee80211_release_reorder_timeout(struct sta_info *sta, int tid);
 /* Spectrum management */
 void ieee80211_process_measurement_req(struct ieee80211_sub_if_data *sdata,
@@ -1146,6 +1212,12 @@ int __ieee80211_suspend(struct ieee80211_hw *hw);
 static inline int __ieee80211_resume(struct ieee80211_hw *hw)
 {
+        struct ieee80211_local *local = hw_to_local(hw);
+        WARN(test_bit(SCAN_HW_SCANNING, &local->scanning),
+                "%s: resume with hardware scan still in progress\n",
+                wiphy_name(hw->wiphy));
        return ieee80211_reconfig(hw_to_local(hw));
 }
 #else
@@ -1208,7 +1280,8 @@ void ieee80211_send_auth(struct ieee80211_sub_if_data *sdata,
                         const u8 *key, u8 key_len, u8 key_idx);
 int ieee80211_build_preq_ies(struct ieee80211_local *local, u8 *buffer,
                             const u8 *ie, size_t ie_len,
-                             enum ieee80211_band band);
+                             enum ieee80211_band band, u32 rate_mask,
+                             u8 channel);
 void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
                              const u8 *ssid, size_t ssid_len,
                              const u8 *ie, size_t ie_len);
diff --git a/net/mac80211/iface.c b/net/mac80211/iface.c
index ebbe264e2b0b..66785739dad3 100644
--- a/net/mac80211/iface.c
+++ b/net/mac80211/iface.c
@@ -94,21 +94,14 @@ static inline int identical_mac_addr_allowed(int type1, int type2)
                         type2 == NL80211_IFTYPE_AP_VLAN));
 }
-static int ieee80211_open(struct net_device *dev)
+static int ieee80211_check_concurrent_iface(struct ieee80211_sub_if_data *sdata,
+                                            enum nl80211_iftype iftype)
 {
-        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
-        struct ieee80211_sub_if_data *nsdata;
        struct ieee80211_local *local = sdata->local;
-        struct sta_info *sta;
+        struct ieee80211_sub_if_data *nsdata;
-        u32 changed = 0;
+        struct net_device *dev = sdata->dev;
-        int res;
-        u32 hw_reconf_flags = 0;
-        u8 null_addr[ETH_ALEN] = {0};
-        /* fail early if user set an invalid address */
+        ASSERT_RTNL();
-        if (compare_ether_addr(dev->dev_addr, null_addr) &&
-            !is_valid_ether_addr(dev->dev_addr))
-                return -EADDRNOTAVAIL;
        /* we hold the RTNL here so can safely walk the list */
        list_for_each_entry(nsdata, &local->interfaces, list) {
@@ -125,7 +118,7 @@ static int ieee80211_open(struct net_device *dev)
                         * belonging to the same hardware. Then, however, we're
                         * faced with having to adopt two different TSF timers...
                         */
-                        if (sdata->vif.type == NL80211_IFTYPE_ADHOC &&
+                        if (iftype == NL80211_IFTYPE_ADHOC &&
                            nsdata->vif.type == NL80211_IFTYPE_ADHOC)
                                return -EBUSY;
@@ -139,19 +132,36 @@ static int ieee80211_open(struct net_device *dev)
                        /*
                         * check whether it may have the same address
                         */
-                        if (!identical_mac_addr_allowed(sdata->vif.type,
+                        if (!identical_mac_addr_allowed(iftype,
                                                        nsdata->vif.type))
                                return -ENOTUNIQ;
                        /*
                         * can only add VLANs to enabled APs
                         */
-                        if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN &&
+                        if (iftype == NL80211_IFTYPE_AP_VLAN &&
                            nsdata->vif.type == NL80211_IFTYPE_AP)
                                sdata->bss = &nsdata->u.ap;
                }
        }
+        return 0;
+}
+/*
+ * NOTE: Be very careful when changing this function, it must NOT return
+ * an error on interface type changes that have been pre-checked, so most
+ * checks should be in ieee80211_check_concurrent_iface.
+ */
+static int ieee80211_do_open(struct net_device *dev, bool coming_up)
+{
+        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+        struct ieee80211_local *local = sdata->local;
+        struct sta_info *sta;
+        u32 changed = 0;
+        int res;
+        u32 hw_reconf_flags = 0;
        switch (sdata->vif.type) {
        case NL80211_IFTYPE_WDS:
                if (!is_valid_ether_addr(sdata->u.wds.remote_addr))
@@ -177,7 +187,9 @@ static int ieee80211_open(struct net_device *dev)
                /* no special treatment */
                break;
        case NL80211_IFTYPE_UNSPECIFIED:
-        case __NL80211_IFTYPE_AFTER_LAST:
+        case NUM_NL80211_IFTYPES:
+        case NL80211_IFTYPE_P2P_CLIENT:
+        case NL80211_IFTYPE_P2P_GO:
                /* cannot happen */
                WARN_ON(1);
                break;
@@ -187,39 +199,30 @@ static int ieee80211_open(struct net_device *dev)
                res = drv_start(local);
                if (res)
                        goto err_del_bss;
+                if (local->ops->napi_poll)
+                        napi_enable(&local->napi);
                /* we're brought up, everything changes */
                hw_reconf_flags = ~0;
                ieee80211_led_radio(local, true);
        }
        /*
-         * Check all interfaces and copy the hopefully now-present
+         * Copy the hopefully now-present MAC address to
-         * MAC address to those that have the special null one.
+         * this interface, if it has the special null one.
         */
-        list_for_each_entry(nsdata, &local->interfaces, list) {
+        if (is_zero_ether_addr(dev->dev_addr)) {
-                struct net_device *ndev = nsdata->dev;
+                memcpy(dev->dev_addr,
+                       local->hw.wiphy->perm_addr,
-                /*
+                       ETH_ALEN);
-                 * No need to check running since we do not allow
+                memcpy(dev->perm_addr, dev->dev_addr, ETH_ALEN);
-                 * it to start up with this invalid address.
-                 */
+                if (!is_valid_ether_addr(dev->dev_addr)) {
-                if (compare_ether_addr(null_addr, ndev->dev_addr) == 0) {
+                        if (!local->open_count)
-                        memcpy(ndev->dev_addr,
+                                drv_stop(local);
-                               local->hw.wiphy->perm_addr,
+                        return -EADDRNOTAVAIL;
-                               ETH_ALEN);
-                        memcpy(ndev->perm_addr, ndev->dev_addr, ETH_ALEN);
                }
        }
-        /*
-         * Validate the MAC address for this device.
-         */
-        if (!is_valid_ether_addr(dev->dev_addr)) {
-                if (!local->open_count)
-                        drv_stop(local);
-                return -EADDRNOTAVAIL;
-        }
        switch (sdata->vif.type) {
        case NL80211_IFTYPE_AP_VLAN:
                /* no need to tell driver */
@@ -253,9 +256,11 @@ static int ieee80211_open(struct net_device *dev)
                netif_carrier_on(dev);
                break;
        default:
-                res = drv_add_interface(local, &sdata->vif);
+                if (coming_up) {
-                if (res)
+                        res = drv_add_interface(local, &sdata->vif);
-                        goto err_stop;
+                        if (res)
+                                goto err_stop;
+                }
                if (ieee80211_vif_is_mesh(&sdata->vif)) {
                        local->fif_other_bss++;
@@ -277,6 +282,8 @@ static int ieee80211_open(struct net_device *dev)
                        netif_carrier_on(dev);
        }
+        set_bit(SDATA_STATE_RUNNING, &sdata->state);
        if (sdata->vif.type == NL80211_IFTYPE_WDS) {
                /* Create STA entry for the WDS peer */
                sta = sta_info_alloc(sdata, sdata->u.wds.remote_addr,
@@ -307,9 +314,13 @@ static int ieee80211_open(struct net_device *dev)
        if (sdata->flags & IEEE80211_SDATA_PROMISC)
                atomic_inc(&local->iff_promiscs);
+        mutex_lock(&local->mtx);
        hw_reconf_flags |= __ieee80211_recalc_idle(local);
+        mutex_unlock(&local->mtx);
+        if (coming_up)
+                local->open_count++;
-        local->open_count++;
        if (hw_reconf_flags) {
                ieee80211_hw_config(local, hw_reconf_flags);
                /*
@@ -334,22 +345,42 @@ static int ieee80211_open(struct net_device *dev)
        sdata->bss = NULL;
        if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
                list_del(&sdata->u.vlan.list);
+        clear_bit(SDATA_STATE_RUNNING, &sdata->state);
        return res;
 }
-static int ieee80211_stop(struct net_device *dev)
+static int ieee80211_open(struct net_device *dev)
 {
        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+        int err;
+        /* fail early if user set an invalid address */
+        if (!is_zero_ether_addr(dev->dev_addr) &&
+            !is_valid_ether_addr(dev->dev_addr))
+                return -EADDRNOTAVAIL;
+        err = ieee80211_check_concurrent_iface(sdata, sdata->vif.type);
+        if (err)
+                return err;
+        return ieee80211_do_open(dev, true);
+}
+static void ieee80211_do_stop(struct ieee80211_sub_if_data *sdata,
+                              bool going_down)
+{
        struct ieee80211_local *local = sdata->local;
        unsigned long flags;
        struct sk_buff *skb, *tmp;
        u32 hw_reconf_flags = 0;
        int i;
+        clear_bit(SDATA_STATE_RUNNING, &sdata->state);
        /*
         * Stop TX on this interface first.
         */
-        netif_tx_stop_all_queues(dev);
+        netif_tx_stop_all_queues(sdata->dev);
        /*
         * Purge work for this interface.
@@ -366,12 +397,9 @@ static int ieee80211_stop(struct net_device *dev)
         * (because if we remove a STA after ops->remove_interface()
         * the driver will have removed the vif info already!)
         *
-         * We could relax this and only unlink the stations from the
+         * This is relevant only in AP, WDS and mesh modes, since in
-         * hash table and list but keep them on a per-sdata list that
+         * all other modes we've already removed all stations when
-         * will be inserted back again when the interface is brought
+         * disconnecting etc.
-         * up again, but I don't currently see a use case for that,
-         * except with WDS which gets a STA entry created when it is
-         * brought up.
         */
        sta_info_flush(local, sdata);
@@ -390,11 +418,12 @@ static int ieee80211_stop(struct net_device *dev)
        if (sdata->vif.type == NL80211_IFTYPE_AP)
                local->fif_pspoll--;
-        netif_addr_lock_bh(dev);
+        netif_addr_lock_bh(sdata->dev);
        spin_lock_bh(&local->filter_lock);
-        __hw_addr_unsync(&local->mc_list, &dev->mc, dev->addr_len);
+        __hw_addr_unsync(&local->mc_list, &sdata->dev->mc,
+                         sdata->dev->addr_len);
        spin_unlock_bh(&local->filter_lock);
-        netif_addr_unlock_bh(dev);
+        netif_addr_unlock_bh(sdata->dev);
        ieee80211_configure_filter(local);
@@ -406,11 +435,21 @@ static int ieee80211_stop(struct net_device *dev)
                struct ieee80211_sub_if_data *vlan, *tmpsdata;
                struct beacon_data *old_beacon = sdata->u.ap.beacon;
+                /* sdata_running will return false, so this will disable */
+                ieee80211_bss_info_change_notify(sdata,
+                                                 BSS_CHANGED_BEACON_ENABLED);
                /* remove beacon */
                rcu_assign_pointer(sdata->u.ap.beacon, NULL);
                synchronize_rcu();
                kfree(old_beacon);
+                /* free all potentially still buffered bcast frames */
+                while ((skb = skb_dequeue(&sdata->u.ap.ps_bc_buf))) {
+                        local->total_ps_buffered--;
+                        dev_kfree_skb(skb);
+                }
                /* down all dependent devices, that is VLANs */
                list_for_each_entry_safe(vlan, tmpsdata, &sdata->u.ap.vlans,
                                         u.vlan.list)
@@ -418,7 +457,8 @@ static int ieee80211_stop(struct net_device *dev)
                WARN_ON(!list_empty(&sdata->u.ap.vlans));
        }
-        local->open_count--;
+        if (going_down)
+                local->open_count--;
        switch (sdata->vif.type) {
        case NL80211_IFTYPE_AP_VLAN:
@@ -450,27 +490,6 @@ static int ieee80211_stop(struct net_device *dev)
                ieee80211_configure_filter(local);
                break;
-        case NL80211_IFTYPE_STATION:
-                del_timer_sync(&sdata->u.mgd.chswitch_timer);
-                del_timer_sync(&sdata->u.mgd.timer);
-                del_timer_sync(&sdata->u.mgd.conn_mon_timer);
-                del_timer_sync(&sdata->u.mgd.bcn_mon_timer);
-                /*
-                 * If any of the timers fired while we waited for it, it will
-                 * have queued its work. Now the work will be running again
-                 * but will not rearm the timer again because it checks
-                 * whether the interface is running, which, at this point,
-                 * it no longer is.
-                 */
-                cancel_work_sync(&sdata->u.mgd.chswitch_work);
-                cancel_work_sync(&sdata->u.mgd.monitor_work);
-                cancel_work_sync(&sdata->u.mgd.beacon_connection_loss_work);
-                /* fall through */
-        case NL80211_IFTYPE_ADHOC:
-                if (sdata->vif.type == NL80211_IFTYPE_ADHOC)
-                        del_timer_sync(&sdata->u.ibss.timer);
-                /* fall through */
        case NL80211_IFTYPE_MESH_POINT:
                if (ieee80211_vif_is_mesh(&sdata->vif)) {
                        /* other_bss and allmulti are always set on mesh
@@ -498,27 +517,34 @@ static int ieee80211_stop(struct net_device *dev)
                        ieee80211_scan_cancel(local);
                /*
-                 * Disable beaconing for AP and mesh, IBSS can't
+                 * Disable beaconing here for mesh only, AP and IBSS
-                 * still be joined to a network at this point.
+                 * are already taken care of.
                 */
-                if (sdata->vif.type == NL80211_IFTYPE_AP ||
+                if (sdata->vif.type == NL80211_IFTYPE_MESH_POINT)
-                    sdata->vif.type == NL80211_IFTYPE_MESH_POINT) {
                        ieee80211_bss_info_change_notify(sdata,
                                BSS_CHANGED_BEACON_ENABLED);
-                }
-                /* free all remaining keys, there shouldn't be any */
+                /*
+                 * Free all remaining keys, there shouldn't be any,
+                 * except maybe group keys in AP more or WDS?
+                 */
                ieee80211_free_keys(sdata);
-                drv_remove_interface(local, &sdata->vif);
+                if (going_down)
+                        drv_remove_interface(local, &sdata->vif);
        }
        sdata->bss = NULL;
+        mutex_lock(&local->mtx);
        hw_reconf_flags |= __ieee80211_recalc_idle(local);
+        mutex_unlock(&local->mtx);
        ieee80211_recalc_ps(local, -1);
        if (local->open_count == 0) {
+                if (local->ops->napi_poll)
+                        napi_disable(&local->napi);
                ieee80211_clear_tx_pending(local);
                ieee80211_stop_device(local);
@@ -541,6 +567,13 @@ static int ieee80211_stop(struct net_device *dev)
                }
        }
        spin_unlock_irqrestore(&local->queue_stop_reason_lock, flags);
+}
+static int ieee80211_stop(struct net_device *dev)
+{
+        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
+        ieee80211_do_stop(sdata, true);
        return 0;
 }
@@ -585,8 +618,6 @@ static void ieee80211_teardown_sdata(struct net_device *dev)
 {
        struct ieee80211_sub_if_data *sdata = IEEE80211_DEV_TO_SUB_IF(dev);
        struct ieee80211_local *local = sdata->local;
-        struct beacon_data *beacon;
-        struct sk_buff *skb;
        int flushed;
        int i;
@@ -599,37 +630,8 @@ static void ieee80211_teardown_sdata(struct net_device *dev)
                __skb_queue_purge(&sdata->fragments[i].skb_list);
        sdata->fragment_next = 0;
-        switch (sdata->vif.type) {
+        if (ieee80211_vif_is_mesh(&sdata->vif))
-        case NL80211_IFTYPE_AP:
+                mesh_rmc_free(sdata);
-                beacon = sdata->u.ap.beacon;
-                rcu_assign_pointer(sdata->u.ap.beacon, NULL);
-                synchronize_rcu();
-                kfree(beacon);
-                while ((skb = skb_dequeue(&sdata->u.ap.ps_bc_buf))) {
-                        local->total_ps_buffered--;
-                        dev_kfree_skb(skb);
-                }
-                break;
-        case NL80211_IFTYPE_MESH_POINT:
-                if (ieee80211_vif_is_mesh(&sdata->vif))
-                        mesh_rmc_free(sdata);
-                break;
-        case NL80211_IFTYPE_ADHOC:
-                if (WARN_ON(sdata->u.ibss.presp))
-                        kfree_skb(sdata->u.ibss.presp);
-                break;
-        case NL80211_IFTYPE_STATION:
-        case NL80211_IFTYPE_WDS:
-        case NL80211_IFTYPE_AP_VLAN:
-        case NL80211_IFTYPE_MONITOR:
-                break;
-        case NL80211_IFTYPE_UNSPECIFIED:
-        case __NL80211_IFTYPE_AFTER_LAST:
-                BUG();
-                break;
-        }
        flushed = sta_info_flush(local, sdata);
        WARN_ON(flushed);
@@ -844,9 +846,13 @@ static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata,
        /* and set some type-dependent values */
        sdata->vif.type = type;
+        sdata->vif.p2p = false;
        sdata->dev->netdev_ops = &ieee80211_dataif_ops;
        sdata->wdev.iftype = type;
+        sdata->control_port_protocol = cpu_to_be16(ETH_P_PAE);
+        sdata->control_port_no_encrypt = false;
        /* only monitor differs */
        sdata->dev->type = ARPHRD_ETHER;
@@ -854,10 +860,20 @@ static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata,
        INIT_WORK(&sdata->work, ieee80211_iface_work);
        switch (type) {
+        case NL80211_IFTYPE_P2P_GO:
+                type = NL80211_IFTYPE_AP;
+                sdata->vif.type = type;
+                sdata->vif.p2p = true;
+                /* fall through */
        case NL80211_IFTYPE_AP:
                skb_queue_head_init(&sdata->u.ap.ps_bc_buf);
                INIT_LIST_HEAD(&sdata->u.ap.vlans);
                break;
+        case NL80211_IFTYPE_P2P_CLIENT:
+                type = NL80211_IFTYPE_STATION;
+                sdata->vif.type = type;
+                sdata->vif.p2p = true;
+                /* fall through */
        case NL80211_IFTYPE_STATION:
                ieee80211_sta_setup_sdata(sdata);
                break;
@@ -878,7 +894,7 @@ static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata,
        case NL80211_IFTYPE_AP_VLAN:
                break;
        case NL80211_IFTYPE_UNSPECIFIED:
-        case __NL80211_IFTYPE_AFTER_LAST:
+        case NUM_NL80211_IFTYPES:
                BUG();
                break;
        }
@@ -886,12 +902,85 @@ static void ieee80211_setup_sdata(struct ieee80211_sub_if_data *sdata,
        ieee80211_debugfs_add_netdev(sdata);
 }
+static int ieee80211_runtime_change_iftype(struct ieee80211_sub_if_data *sdata,
+                                           enum nl80211_iftype type)
+{
+        struct ieee80211_local *local = sdata->local;
+        int ret, err;
+        enum nl80211_iftype internal_type = type;
+        bool p2p = false;
+        ASSERT_RTNL();
+        if (!local->ops->change_interface)
+                return -EBUSY;
+        switch (sdata->vif.type) {
+        case NL80211_IFTYPE_AP:
+        case NL80211_IFTYPE_STATION:
+        case NL80211_IFTYPE_ADHOC:
+                /*
+                 * Could maybe also all others here?
+                 * Just not sure how that interacts
+                 * with the RX/config path e.g. for
+                 * mesh.
+                 */
+                break;
+        default:
+                return -EBUSY;
+        }
+        switch (type) {
+        case NL80211_IFTYPE_AP:
+        case NL80211_IFTYPE_STATION:
+        case NL80211_IFTYPE_ADHOC:
+                /*
+                 * Could probably support everything
+                 * but WDS here (WDS do_open can fail
+                 * under memory pressure, which this
+                 * code isn't prepared to handle).
+                 */
+                break;
+        case NL80211_IFTYPE_P2P_CLIENT:
+                p2p = true;
+                internal_type = NL80211_IFTYPE_STATION;
+                break;
+        case NL80211_IFTYPE_P2P_GO:
+                p2p = true;
+                internal_type = NL80211_IFTYPE_AP;
+                break;
+        default:
+                return -EBUSY;
+        }
+        ret = ieee80211_check_concurrent_iface(sdata, internal_type);
+        if (ret)
+                return ret;
+        ieee80211_do_stop(sdata, false);
+        ieee80211_teardown_sdata(sdata->dev);
+        ret = drv_change_interface(local, sdata, internal_type, p2p);
+        if (ret)
+                type = sdata->vif.type;
+        ieee80211_setup_sdata(sdata, type);
+        err = ieee80211_do_open(sdata->dev, false);
+        WARN(err, "type change: do_open returned %d", err);
+        return ret;
+}
 int ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata,
                             enum nl80211_iftype type)
 {
+        int ret;
        ASSERT_RTNL();
-        if (type == sdata->vif.type)
+        if (type == ieee80211_vif_type_p2p(&sdata->vif))
                return 0;
        /* Setting ad-hoc mode on non-IBSS channel is not supported. */
@@ -899,18 +988,15 @@ int ieee80211_if_change_type(struct ieee80211_sub_if_data *sdata,
            type == NL80211_IFTYPE_ADHOC)
                return -EOPNOTSUPP;
-        /*
+        if (ieee80211_sdata_running(sdata)) {
-         * We could, here, on changes between IBSS/STA/MESH modes,
+                ret = ieee80211_runtime_change_iftype(sdata, type);
-         * invoke an MLME function instead that disassociates etc.
+                if (ret)
-         * and goes into the requested mode.
+                        return ret;
-         */
+        } else {
+                /* Purge and reset type-dependent state. */
-        if (ieee80211_sdata_running(sdata))
+                ieee80211_teardown_sdata(sdata->dev);
-                return -EBUSY;
+                ieee80211_setup_sdata(sdata, type);
+        }
-        /* Purge and reset type-dependent state. */
-        ieee80211_teardown_sdata(sdata->dev);
-        ieee80211_setup_sdata(sdata, type);
        /* reset some values that shouldn't be kept across type changes */
        sdata->vif.bss_conf.basic_rates =
@@ -1167,8 +1253,7 @@ static u32 ieee80211_idle_off(struct ieee80211_local *local,
                return 0;
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: device no longer idle - %s\n",
+        wiphy_debug(local->hw.wiphy, "device no longer idle - %s\n", reason);
-               wiphy_name(local->hw.wiphy), reason);
 #endif
        local->hw.conf.flags &= ~IEEE80211_CONF_IDLE;
@@ -1181,8 +1266,7 @@ static u32 ieee80211_idle_on(struct ieee80211_local *local)
                return 0;
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: device now idle\n",
+        wiphy_debug(local->hw.wiphy, "device now idle\n");
-               wiphy_name(local->hw.wiphy));
 #endif
        drv_flush(local, false);
@@ -1195,28 +1279,61 @@ u32 __ieee80211_recalc_idle(struct ieee80211_local *local)
 {
        struct ieee80211_sub_if_data *sdata;
        int count = 0;
+        bool working = false, scanning = false;
+        struct ieee80211_work *wk;
-        if (!list_empty(&local->work_list))
+#ifdef CONFIG_PROVE_LOCKING
-                return ieee80211_idle_off(local, "working");
+        WARN_ON(debug_locks && !lockdep_rtnl_is_held() &&
+                !lockdep_is_held(&local->iflist_mtx));
-        if (local->scanning)
+#endif
-                return ieee80211_idle_off(local, "scanning");
+        lockdep_assert_held(&local->mtx);
        list_for_each_entry(sdata, &local->interfaces, list) {
-                if (!ieee80211_sdata_running(sdata))
+                if (!ieee80211_sdata_running(sdata)) {
+                        sdata->vif.bss_conf.idle = true;
                        continue;
+                }
+                sdata->old_idle = sdata->vif.bss_conf.idle;
                /* do not count disabled managed interfaces */
                if (sdata->vif.type == NL80211_IFTYPE_STATION &&
-                    !sdata->u.mgd.associated)
+                    !sdata->u.mgd.associated) {
+                        sdata->vif.bss_conf.idle = true;
                        continue;
+                }
                /* do not count unused IBSS interfaces */
                if (sdata->vif.type == NL80211_IFTYPE_ADHOC &&
-                    !sdata->u.ibss.ssid_len)
+                    !sdata->u.ibss.ssid_len) {
+                        sdata->vif.bss_conf.idle = true;
                        continue;
+                }
                /* count everything else */
                count++;
        }
+        list_for_each_entry(wk, &local->work_list, list) {
+                working = true;
+                wk->sdata->vif.bss_conf.idle = false;
+        }
+        if (local->scan_sdata) {
+                scanning = true;
+                local->scan_sdata->vif.bss_conf.idle = false;
+        }
+        list_for_each_entry(sdata, &local->interfaces, list) {
+                if (sdata->old_idle == sdata->vif.bss_conf.idle)
+                        continue;
+                if (!ieee80211_sdata_running(sdata))
+                        continue;
+                ieee80211_bss_info_change_notify(sdata, BSS_CHANGED_IDLE);
+        }
+        if (working)
+                return ieee80211_idle_off(local, "working");
+        if (scanning)
+                return ieee80211_idle_off(local, "scanning");
        if (!count)
                return ieee80211_idle_on(local);
        else
diff --git a/net/mac80211/key.c b/net/mac80211/key.c
index 1b9d87ed143a..6a63d1abd14d 100644
--- a/net/mac80211/key.c
+++ b/net/mac80211/key.c
@@ -49,7 +49,7 @@ static const u8 bcast_addr[ETH_ALEN] = { 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF };
 static void assert_key_lock(struct ieee80211_local *local)
 {
-        WARN_ON(!mutex_is_locked(&local->key_mtx));
+        lockdep_assert_held(&local->key_mtx);
 }
 static struct ieee80211_sta *get_sta_for_key(struct ieee80211_key *key)
@@ -60,7 +60,7 @@ static struct ieee80211_sta *get_sta_for_key(struct ieee80211_key *key)
        return NULL;
 }
-static void ieee80211_key_enable_hw_accel(struct ieee80211_key *key)
+static int ieee80211_key_enable_hw_accel(struct ieee80211_key *key)
 {
        struct ieee80211_sub_if_data *sdata;
        struct ieee80211_sta *sta;
@@ -68,8 +68,10 @@ static void ieee80211_key_enable_hw_accel(struct ieee80211_key *key)
        might_sleep();
-        if (!key->local->ops->set_key)
+        if (!key->local->ops->set_key) {
-                return;
+                ret = -EOPNOTSUPP;
+                goto out_unsupported;
+        }
        assert_key_lock(key->local);
@@ -87,10 +89,27 @@ static void ieee80211_key_enable_hw_accel(struct ieee80211_key *key)
                key->flags |= KEY_FLAG_UPLOADED_TO_HARDWARE;
        if (ret && ret != -ENOSPC && ret != -EOPNOTSUPP)
-                printk(KERN_ERR "mac80211-%s: failed to set key "
+                wiphy_err(key->local->hw.wiphy,
-                       "(%d, %pM) to hardware (%d)\n",
+                          "failed to set key (%d, %pM) to hardware (%d)\n",
-                       wiphy_name(key->local->hw.wiphy),
+                          key->conf.keyidx, sta ? sta->addr : bcast_addr, ret);
-                       key->conf.keyidx, sta ? sta->addr : bcast_addr, ret);
+out_unsupported:
+        if (ret) {
+                switch (key->conf.cipher) {
+                case WLAN_CIPHER_SUITE_WEP40:
+                case WLAN_CIPHER_SUITE_WEP104:
+                case WLAN_CIPHER_SUITE_TKIP:
+                case WLAN_CIPHER_SUITE_CCMP:
+                case WLAN_CIPHER_SUITE_AES_CMAC:
+                        /* all of these we can do in software */
+                        ret = 0;
+                        break;
+                default:
+                        ret = -EINVAL;
+                }
+        }
+        return ret;
 }
 static void ieee80211_key_disable_hw_accel(struct ieee80211_key *key)
@@ -121,10 +140,9 @@ static void ieee80211_key_disable_hw_accel(struct ieee80211_key *key)
                          sta, &key->conf);
        if (ret)
-                printk(KERN_ERR "mac80211-%s: failed to remove key "
+                wiphy_err(key->local->hw.wiphy,
-                       "(%d, %pM) from hardware (%d)\n",
+                          "failed to remove key (%d, %pM) from hardware (%d)\n",
-                       wiphy_name(key->local->hw.wiphy),
+                          key->conf.keyidx, sta ? sta->addr : bcast_addr, ret);
-                       key->conf.keyidx, sta ? sta->addr : bcast_addr, ret);
        key->flags &= ~KEY_FLAG_UPLOADED_TO_HARDWARE;
 }
@@ -227,20 +245,18 @@ static void __ieee80211_key_replace(struct ieee80211_sub_if_data *sdata,
        }
 }
-struct ieee80211_key *ieee80211_key_alloc(enum ieee80211_key_alg alg,
+struct ieee80211_key *ieee80211_key_alloc(u32 cipher, int idx, size_t key_len,
-                                          int idx,
-                                          size_t key_len,
                                          const u8 *key_data,
                                          size_t seq_len, const u8 *seq)
 {
        struct ieee80211_key *key;
-        int i, j;
+        int i, j, err;
        BUG_ON(idx < 0 || idx >= NUM_DEFAULT_KEYS + NUM_DEFAULT_MGMT_KEYS);
        key = kzalloc(sizeof(struct ieee80211_key) + key_len, GFP_KERNEL);
        if (!key)
-                return NULL;
+                return ERR_PTR(-ENOMEM);
        /*
         * Default to software encryption; we'll later upload the
@@ -249,15 +265,16 @@ struct ieee80211_key *ieee80211_key_alloc(enum ieee80211_key_alg alg,
        key->conf.flags = 0;
        key->flags = 0;
-        key->conf.alg = alg;
+        key->conf.cipher = cipher;
        key->conf.keyidx = idx;
        key->conf.keylen = key_len;
-        switch (alg) {
+        switch (cipher) {
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_WEP40:
+        case WLAN_CIPHER_SUITE_WEP104:
                key->conf.iv_len = WEP_IV_LEN;
                key->conf.icv_len = WEP_ICV_LEN;
                break;
-        case ALG_TKIP:
+        case WLAN_CIPHER_SUITE_TKIP:
                key->conf.iv_len = TKIP_IV_LEN;
                key->conf.icv_len = TKIP_ICV_LEN;
                if (seq) {
@@ -269,7 +286,7 @@ struct ieee80211_key *ieee80211_key_alloc(enum ieee80211_key_alg alg,
                        }
                }
                break;
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                key->conf.iv_len = CCMP_HDR_LEN;
                key->conf.icv_len = CCMP_MIC_LEN;
                if (seq) {
@@ -278,42 +295,38 @@ struct ieee80211_key *ieee80211_key_alloc(enum ieee80211_key_alg alg,
                                        key->u.ccmp.rx_pn[i][j] =
                                                seq[CCMP_PN_LEN - j - 1];
                }
-                break;
-        case ALG_AES_CMAC:
-                key->conf.iv_len = 0;
-                key->conf.icv_len = sizeof(struct ieee80211_mmie);
-                if (seq)
-                        for (j = 0; j < 6; j++)
-                                key->u.aes_cmac.rx_pn[j] = seq[6 - j - 1];
-                break;
-        }
-        memcpy(key->conf.key, key_data, key_len);
-        INIT_LIST_HEAD(&key->list);
-        if (alg == ALG_CCMP) {
                /*
                 * Initialize AES key state here as an optimization so that
                 * it does not need to be initialized for every packet.
                 */
                key->u.ccmp.tfm = ieee80211_aes_key_setup_encrypt(key_data);
-                if (!key->u.ccmp.tfm) {
+                if (IS_ERR(key->u.ccmp.tfm)) {
+                        err = PTR_ERR(key->u.ccmp.tfm);
                        kfree(key);
-                        return NULL;
+                        key = ERR_PTR(err);
                }
-        }
+                break;
+        case WLAN_CIPHER_SUITE_AES_CMAC:
-        if (alg == ALG_AES_CMAC) {
+                key->conf.iv_len = 0;
+                key->conf.icv_len = sizeof(struct ieee80211_mmie);
+                if (seq)
+                        for (j = 0; j < 6; j++)
+                                key->u.aes_cmac.rx_pn[j] = seq[6 - j - 1];
                /*
                 * Initialize AES key state here as an optimization so that
                 * it does not need to be initialized for every packet.
                 */
                key->u.aes_cmac.tfm =
                        ieee80211_aes_cmac_key_setup(key_data);
-                if (!key->u.aes_cmac.tfm) {
+                if (IS_ERR(key->u.aes_cmac.tfm)) {
+                        err = PTR_ERR(key->u.aes_cmac.tfm);
                        kfree(key);
-                        return NULL;
+                        key = ERR_PTR(err);
                }
+                break;
        }
+        memcpy(key->conf.key, key_data, key_len);
+        INIT_LIST_HEAD(&key->list);
        return key;
 }
@@ -326,9 +339,9 @@ static void __ieee80211_key_destroy(struct ieee80211_key *key)
        if (key->local)
                ieee80211_key_disable_hw_accel(key);
-        if (key->conf.alg == ALG_CCMP)
+        if (key->conf.cipher == WLAN_CIPHER_SUITE_CCMP)
                ieee80211_aes_key_free(key->u.ccmp.tfm);
-        if (key->conf.alg == ALG_AES_CMAC)
+        if (key->conf.cipher == WLAN_CIPHER_SUITE_AES_CMAC)
                ieee80211_aes_cmac_key_free(key->u.aes_cmac.tfm);
        if (key->local)
                ieee80211_debugfs_key_remove(key);
@@ -336,12 +349,12 @@ static void __ieee80211_key_destroy(struct ieee80211_key *key)
        kfree(key);
 }
-void ieee80211_key_link(struct ieee80211_key *key,
+int ieee80211_key_link(struct ieee80211_key *key,
-                        struct ieee80211_sub_if_data *sdata,
+                       struct ieee80211_sub_if_data *sdata,
-                        struct sta_info *sta)
+                       struct sta_info *sta)
 {
        struct ieee80211_key *old_key;
-        int idx;
+        int idx, ret;
        BUG_ON(!sdata);
        BUG_ON(!key);
@@ -396,9 +409,11 @@ void ieee80211_key_link(struct ieee80211_key *key,
        ieee80211_debugfs_key_add(key);
-        ieee80211_key_enable_hw_accel(key);
+        ret = ieee80211_key_enable_hw_accel(key);
        mutex_unlock(&sdata->local->key_mtx);
+        return ret;
 }
 static void __ieee80211_key_free(struct ieee80211_key *key)
diff --git a/net/mac80211/key.h b/net/mac80211/key.h
index b665bbb7a471..cb9a4a65cc68 100644
--- a/net/mac80211/key.h
+++ b/net/mac80211/key.h
@@ -123,18 +123,16 @@ struct ieee80211_key {
        struct ieee80211_key_conf conf;
 };
-struct ieee80211_key *ieee80211_key_alloc(enum ieee80211_key_alg alg,
+struct ieee80211_key *ieee80211_key_alloc(u32 cipher, int idx, size_t key_len,
-                                          int idx,
-                                          size_t key_len,
                                          const u8 *key_data,
                                          size_t seq_len, const u8 *seq);
 /*
 * Insert a key into data structures (sdata, sta if necessary)
 * to make it used, free old key.
 */
-void ieee80211_key_link(struct ieee80211_key *key,
+int __must_check ieee80211_key_link(struct ieee80211_key *key,
-                        struct ieee80211_sub_if_data *sdata,
+                                    struct ieee80211_sub_if_data *sdata,
-                        struct sta_info *sta);
+                                    struct sta_info *sta);
 void ieee80211_key_free(struct ieee80211_local *local,
                        struct ieee80211_key *key);
 void ieee80211_set_default_key(struct ieee80211_sub_if_data *sdata, int idx);
diff --git a/net/mac80211/main.c b/net/mac80211/main.c
index ded5c3843e06..db341a99c7c7 100644
--- a/net/mac80211/main.c
+++ b/net/mac80211/main.c
@@ -99,16 +99,19 @@ int ieee80211_hw_config(struct ieee80211_local *local, u32 changed)
        int ret = 0;
        int power;
        enum nl80211_channel_type channel_type;
+        u32 offchannel_flag;
        might_sleep();
        scan_chan = local->scan_channel;
+        offchannel_flag = local->hw.conf.flags & IEEE80211_CONF_OFFCHANNEL;
        if (scan_chan) {
                chan = scan_chan;
                channel_type = NL80211_CHAN_NO_HT;
                local->hw.conf.flags |= IEEE80211_CONF_OFFCHANNEL;
-        } else if (local->tmp_channel) {
+        } else if (local->tmp_channel &&
+                   local->oper_channel != local->tmp_channel) {
                chan = scan_chan = local->tmp_channel;
                channel_type = local->tmp_channel_type;
                local->hw.conf.flags |= IEEE80211_CONF_OFFCHANNEL;
@@ -117,8 +120,9 @@ int ieee80211_hw_config(struct ieee80211_local *local, u32 changed)
                channel_type = local->_oper_channel_type;
                local->hw.conf.flags &= ~IEEE80211_CONF_OFFCHANNEL;
        }
+        offchannel_flag ^= local->hw.conf.flags & IEEE80211_CONF_OFFCHANNEL;
-        if (chan != local->hw.conf.channel ||
+        if (offchannel_flag || chan != local->hw.conf.channel ||
            channel_type != local->hw.conf.channel_type) {
                local->hw.conf.channel = chan;
                local->hw.conf.channel_type = channel_type;
@@ -302,7 +306,16 @@ void ieee80211_restart_hw(struct ieee80211_hw *hw)
        trace_api_restart_hw(local);
-        /* use this reason, __ieee80211_resume will unblock it */
+        /* wait for scan work complete */
+        flush_workqueue(local->workqueue);
+        WARN(test_bit(SCAN_HW_SCANNING, &local->scanning),
+                "%s called with hardware scan in progress\n", __func__);
+        if (unlikely(test_bit(SCAN_SW_SCANNING, &local->scanning)))
+                ieee80211_scan_cancel(local);
+        /* use this reason, ieee80211_reconfig will unblock it */
        ieee80211_stop_queues_by_reason(hw,
                IEEE80211_QUEUE_STOP_REASON_SUSPEND);
@@ -336,9 +349,6 @@ static int ieee80211_ifa_changed(struct notifier_block *nb,
        struct ieee80211_if_managed *ifmgd;
        int c = 0;
-        if (!netif_running(ndev))
-                return NOTIFY_DONE;
        /* Make sure it's our interface that got changed */
        if (!wdev)
                return NOTIFY_DONE;
@@ -349,11 +359,14 @@ static int ieee80211_ifa_changed(struct notifier_block *nb,
        sdata = IEEE80211_DEV_TO_SUB_IF(ndev);
        bss_conf = &sdata->vif.bss_conf;
+        if (!ieee80211_sdata_running(sdata))
+                return NOTIFY_DONE;
        /* ARP filtering is only supported in managed mode */
        if (sdata->vif.type != NL80211_IFTYPE_STATION)
                return NOTIFY_DONE;
-        idev = sdata->dev->ip_ptr;
+        idev = __in_dev_get_rtnl(sdata->dev);
        if (!idev)
                return NOTIFY_DONE;
@@ -390,6 +403,80 @@ static int ieee80211_ifa_changed(struct notifier_block *nb,
 }
 #endif
+static int ieee80211_napi_poll(struct napi_struct *napi, int budget)
+{
+        struct ieee80211_local *local =
+                container_of(napi, struct ieee80211_local, napi);
+        return local->ops->napi_poll(&local->hw, budget);
+}
+void ieee80211_napi_schedule(struct ieee80211_hw *hw)
+{
+        struct ieee80211_local *local = hw_to_local(hw);
+        napi_schedule(&local->napi);
+}
+EXPORT_SYMBOL(ieee80211_napi_schedule);
+void ieee80211_napi_complete(struct ieee80211_hw *hw)
+{
+        struct ieee80211_local *local = hw_to_local(hw);
+        napi_complete(&local->napi);
+}
+EXPORT_SYMBOL(ieee80211_napi_complete);
+/* There isn't a lot of sense in it, but you can transmit anything you like */
+static const struct ieee80211_txrx_stypes
+ieee80211_default_mgmt_stypes[NUM_NL80211_IFTYPES] = {
+        [NL80211_IFTYPE_ADHOC] = {
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ACTION >> 4),
+        },
+        [NL80211_IFTYPE_STATION] = {
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ACTION >> 4) |
+                        BIT(IEEE80211_STYPE_PROBE_REQ >> 4),
+        },
+        [NL80211_IFTYPE_AP] = {
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_DISASSOC >> 4) |
+                        BIT(IEEE80211_STYPE_AUTH >> 4) |
+                        BIT(IEEE80211_STYPE_DEAUTH >> 4) |
+                        BIT(IEEE80211_STYPE_ACTION >> 4),
+        },
+        [NL80211_IFTYPE_AP_VLAN] = {
+                /* copy AP */
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_DISASSOC >> 4) |
+                        BIT(IEEE80211_STYPE_AUTH >> 4) |
+                        BIT(IEEE80211_STYPE_DEAUTH >> 4) |
+                        BIT(IEEE80211_STYPE_ACTION >> 4),
+        },
+        [NL80211_IFTYPE_P2P_CLIENT] = {
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ACTION >> 4) |
+                        BIT(IEEE80211_STYPE_PROBE_REQ >> 4),
+        },
+        [NL80211_IFTYPE_P2P_GO] = {
+                .tx = 0xffff,
+                .rx = BIT(IEEE80211_STYPE_ASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_REASSOC_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_PROBE_REQ >> 4) |
+                        BIT(IEEE80211_STYPE_DISASSOC >> 4) |
+                        BIT(IEEE80211_STYPE_AUTH >> 4) |
+                        BIT(IEEE80211_STYPE_DEAUTH >> 4) |
+                        BIT(IEEE80211_STYPE_ACTION >> 4),
+        },
+};
 struct ieee80211_hw *ieee80211_alloc_hw(size_t priv_data_len,
                                        const struct ieee80211_ops *ops)
 {
@@ -419,6 +506,8 @@ struct ieee80211_hw *ieee80211_alloc_hw(size_t priv_data_len,
        if (!wiphy)
                return NULL;
+        wiphy->mgmt_stypes = ieee80211_default_mgmt_stypes;
        wiphy->flags |= WIPHY_FLAG_NETNS_OK |
                        WIPHY_FLAG_4ADDR_AP |
                        WIPHY_FLAG_4ADDR_STATION;
@@ -455,7 +544,7 @@ struct ieee80211_hw *ieee80211_alloc_hw(size_t priv_data_len,
        __hw_addr_init(&local->mc_list);
        mutex_init(&local->iflist_mtx);
-        mutex_init(&local->scan_mtx);
+        mutex_init(&local->mtx);
        mutex_init(&local->key_mtx);
        spin_lock_init(&local->filter_lock);
@@ -494,6 +583,9 @@ struct ieee80211_hw *ieee80211_alloc_hw(size_t priv_data_len,
        skb_queue_head_init(&local->skb_queue);
        skb_queue_head_init(&local->skb_queue_unreliable);
+        /* init dummy netdev for use w/ NAPI */
+        init_dummy_netdev(&local->napi_dev);
        return local_to_hw(local);
 }
 EXPORT_SYMBOL(ieee80211_alloc_hw);
@@ -506,6 +598,7 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
        int channels, max_bitrates;
        bool supp_ht;
        static const u32 cipher_suites[] = {
+                /* keep WEP first, it may be removed below */
                WLAN_CIPHER_SUITE_WEP40,
                WLAN_CIPHER_SUITE_WEP104,
                WLAN_CIPHER_SUITE_TKIP,
@@ -554,6 +647,14 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
        /* mac80211 always supports monitor */
        local->hw.wiphy->interface_modes |= BIT(NL80211_IFTYPE_MONITOR);
+#ifndef CONFIG_MAC80211_MESH
+        /* mesh depends on Kconfig, but drivers should set it if they want */
+        local->hw.wiphy->interface_modes &= ~BIT(NL80211_IFTYPE_MESH_POINT);
+#endif
+        /* mac80211 supports control port protocol changing */
+        local->hw.wiphy->flags |= WIPHY_FLAG_CONTROL_PORT_PROTOCOL;
        if (local->hw.flags & IEEE80211_HW_SIGNAL_DBM)
                local->hw.wiphy->signal_type = CFG80211_SIGNAL_TYPE_MBM;
        else if (local->hw.flags & IEEE80211_HW_SIGNAL_UNSPEC)
@@ -589,10 +690,41 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
        if (local->hw.wiphy->max_scan_ie_len)
                local->hw.wiphy->max_scan_ie_len -= local->scan_ies_len;
-        local->hw.wiphy->cipher_suites = cipher_suites;
+        /* Set up cipher suites unless driver already did */
-        local->hw.wiphy->n_cipher_suites = ARRAY_SIZE(cipher_suites);
+        if (!local->hw.wiphy->cipher_suites) {
-        if (!(local->hw.flags & IEEE80211_HW_MFP_CAPABLE))
+                local->hw.wiphy->cipher_suites = cipher_suites;
-                local->hw.wiphy->n_cipher_suites--;
+                local->hw.wiphy->n_cipher_suites = ARRAY_SIZE(cipher_suites);
+                if (!(local->hw.flags & IEEE80211_HW_MFP_CAPABLE))
+                        local->hw.wiphy->n_cipher_suites--;
+        }
+        if (IS_ERR(local->wep_tx_tfm) || IS_ERR(local->wep_rx_tfm)) {
+                if (local->hw.wiphy->cipher_suites == cipher_suites) {
+                        local->hw.wiphy->cipher_suites += 2;
+                        local->hw.wiphy->n_cipher_suites -= 2;
+                } else {
+                        u32 *suites;
+                        int r, w = 0;
+                        /* Filter out WEP */
+                        suites = kmemdup(
+                                local->hw.wiphy->cipher_suites,
+                                sizeof(u32) * local->hw.wiphy->n_cipher_suites,
+                                GFP_KERNEL);
+                        if (!suites)
+                                return -ENOMEM;
+                        for (r = 0; r < local->hw.wiphy->n_cipher_suites; r++) {
+                                u32 suite = local->hw.wiphy->cipher_suites[r];
+                                if (suite == WLAN_CIPHER_SUITE_WEP40 ||
+                                    suite == WLAN_CIPHER_SUITE_WEP104)
+                                        continue;
+                                suites[w++] = suite;
+                        }
+                        local->hw.wiphy->cipher_suites = suites;
+                        local->hw.wiphy->n_cipher_suites = w;
+                        local->wiphy_ciphers_allocated = true;
+                }
+        }
        result = wiphy_register(local->hw.wiphy);
        if (result < 0)
@@ -641,16 +773,16 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
        result = ieee80211_wep_init(local);
        if (result < 0)
-                printk(KERN_DEBUG "%s: Failed to initialize wep: %d\n",
+                wiphy_debug(local->hw.wiphy, "Failed to initialize wep: %d\n",
-                       wiphy_name(local->hw.wiphy), result);
+                            result);
        rtnl_lock();
        result = ieee80211_init_rate_ctrl_alg(local,
                                              hw->rate_control_algorithm);
        if (result < 0) {
-                printk(KERN_DEBUG "%s: Failed to initialize rate control "
+                wiphy_debug(local->hw.wiphy,
-                       "algorithm\n", wiphy_name(local->hw.wiphy));
+                            "Failed to initialize rate control algorithm\n");
                goto fail_rate;
        }
@@ -659,8 +791,8 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
                result = ieee80211_if_add(local, "wlan%d", NULL,
                                          NL80211_IFTYPE_STATION, NULL);
                if (result)
-                        printk(KERN_WARNING "%s: Failed to add default virtual iface\n",
+                        wiphy_warn(local->hw.wiphy,
-                               wiphy_name(local->hw.wiphy));
+                                   "Failed to add default virtual iface\n");
        }
        rtnl_unlock();
@@ -683,6 +815,9 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
                goto fail_ifa;
 #endif
+        netif_napi_add(&local->napi_dev, &local->napi, ieee80211_napi_poll,
+                        local->hw.napi_weight);
        return 0;
 #ifdef CONFIG_INET
@@ -703,6 +838,8 @@ int ieee80211_register_hw(struct ieee80211_hw *hw)
 fail_workqueue:
        wiphy_unregister(local->hw.wiphy);
 fail_wiphy_register:
+        if (local->wiphy_ciphers_allocated)
+                kfree(local->hw.wiphy->cipher_suites);
        kfree(local->int_scan_req);
        return result;
 }
@@ -738,6 +875,7 @@ void ieee80211_unregister_hw(struct ieee80211_hw *hw)
         */
        del_timer_sync(&local->work_timer);
+        cancel_work_sync(&local->restart_work);
        cancel_work_sync(&local->reconfig_filter);
        ieee80211_clear_tx_pending(local);
@@ -746,8 +884,7 @@ void ieee80211_unregister_hw(struct ieee80211_hw *hw)
        if (skb_queue_len(&local->skb_queue) ||
            skb_queue_len(&local->skb_queue_unreliable))
-                printk(KERN_WARNING "%s: skb_queue not empty\n",
+                wiphy_warn(local->hw.wiphy, "skb_queue not empty\n");
-                       wiphy_name(local->hw.wiphy));
        skb_queue_purge(&local->skb_queue);
        skb_queue_purge(&local->skb_queue_unreliable);
@@ -764,7 +901,10 @@ void ieee80211_free_hw(struct ieee80211_hw *hw)
        struct ieee80211_local *local = hw_to_local(hw);
        mutex_destroy(&local->iflist_mtx);
-        mutex_destroy(&local->scan_mtx);
+        mutex_destroy(&local->mtx);
+        if (local->wiphy_ciphers_allocated)
+                kfree(local->hw.wiphy->cipher_suites);
        wiphy_free(local->hw.wiphy);
 }
diff --git a/net/mac80211/mlme.c b/net/mac80211/mlme.c
index b6c163ac22da..77913a15f537 100644
--- a/net/mac80211/mlme.c
+++ b/net/mac80211/mlme.c
@@ -54,6 +54,12 @@
 */
 #define IEEE80211_SIGNAL_AVE_WEIGHT     3
+/*
+ * How many Beacon frames need to have been used in average signal strength
+ * before starting to indicate signal change events.
+ */
+#define IEEE80211_SIGNAL_AVE_MIN_COUNT  4
 #define TMR_RUNNING_TIMER       0
 #define TMR_RUNNING_CHANSW      1
@@ -86,7 +92,7 @@ enum rx_mgmt_action {
 /* utils */
 static inline void ASSERT_MGD_MTX(struct ieee80211_if_managed *ifmgd)
 {
-        WARN_ON(!mutex_is_locked(&ifmgd->mtx));
+        lockdep_assert_held(&ifmgd->mtx);
 }
 /*
@@ -109,7 +115,7 @@ static void run_again(struct ieee80211_if_managed *ifmgd,
                mod_timer(&ifmgd->timer, timeout);
 }
-static void mod_beacon_timer(struct ieee80211_sub_if_data *sdata)
+void ieee80211_sta_reset_beacon_monitor(struct ieee80211_sub_if_data *sdata)
 {
        if (sdata->local->hw.flags & IEEE80211_HW_BEACON_FILTER)
                return;
@@ -118,6 +124,19 @@ static void mod_beacon_timer(struct ieee80211_sub_if_data *sdata)
                  round_jiffies_up(jiffies + IEEE80211_BEACON_LOSS_TIME));
 }
+void ieee80211_sta_reset_conn_monitor(struct ieee80211_sub_if_data *sdata)
+{
+        struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
+        if (sdata->local->hw.flags & IEEE80211_HW_CONNECTION_MONITOR)
+                return;
+        mod_timer(&sdata->u.mgd.conn_mon_timer,
+                  round_jiffies_up(jiffies + IEEE80211_CONNECTION_IDLE_TIME));
+        ifmgd->probe_send_count = 0;
+}
 static int ecw2cw(int ecw)
 {
        return (1 << ecw) - 1;
@@ -778,16 +797,17 @@ static void ieee80211_sta_wmm_params(struct ieee80211_local *local,
                params.uapsd = uapsd;
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-                printk(KERN_DEBUG "%s: WMM queue=%d aci=%d acm=%d aifs=%d "
+                wiphy_debug(local->hw.wiphy,
-                       "cWmin=%d cWmax=%d txop=%d uapsd=%d\n",
+                            "WMM queue=%d aci=%d acm=%d aifs=%d "
-                       wiphy_name(local->hw.wiphy), queue, aci, acm,
+                            "cWmin=%d cWmax=%d txop=%d uapsd=%d\n",
-                       params.aifs, params.cw_min, params.cw_max, params.txop,
+                            queue, aci, acm,
-                       params.uapsd);
+                            params.aifs, params.cw_min, params.cw_max,
+                            params.txop, params.uapsd);
 #endif
                if (drv_conf_tx(local, queue, &params))
-                        printk(KERN_DEBUG "%s: failed to set TX queue "
+                        wiphy_debug(local->hw.wiphy,
-                               "parameters for queue %d\n",
+                                    "failed to set TX queue parameters for queue %d\n",
-                               wiphy_name(local->hw.wiphy), queue);
+                                    queue);
        }
        /* enable WMM or activate new settings */
@@ -860,14 +880,6 @@ static void ieee80211_set_associated(struct ieee80211_sub_if_data *sdata,
        sdata->u.mgd.flags &= ~(IEEE80211_STA_CONNECTION_POLL |
                                IEEE80211_STA_BEACON_POLL);
-        /*
-         * Always handle WMM once after association regardless
-         * of the first value the AP uses. Setting -1 here has
-         * that effect because the AP values is an unsigned
-         * 4-bit value.
-         */
-        sdata->u.mgd.wmm_last_param_set = -1;
        ieee80211_led_assoc(local, 1);
        if (local->hw.flags & IEEE80211_HW_NEED_DTIM_PERIOD)
@@ -990,6 +1002,11 @@ static void ieee80211_set_disassoc(struct ieee80211_sub_if_data *sdata,
        if (remove_sta)
                sta_info_destroy_addr(sdata, bssid);
+        del_timer_sync(&sdata->u.mgd.conn_mon_timer);
+        del_timer_sync(&sdata->u.mgd.bcn_mon_timer);
+        del_timer_sync(&sdata->u.mgd.timer);
+        del_timer_sync(&sdata->u.mgd.chswitch_timer);
 }
 void ieee80211_sta_rx_notify(struct ieee80211_sub_if_data *sdata,
@@ -1006,21 +1023,26 @@ void ieee80211_sta_rx_notify(struct ieee80211_sub_if_data *sdata,
        if (is_multicast_ether_addr(hdr->addr1))
                return;
-        if (sdata->local->hw.flags & IEEE80211_HW_CONNECTION_MONITOR)
+        ieee80211_sta_reset_conn_monitor(sdata);
-                return;
-        mod_timer(&sdata->u.mgd.conn_mon_timer,
-                  round_jiffies_up(jiffies + IEEE80211_CONNECTION_IDLE_TIME));
 }
 static void ieee80211_mgd_probe_ap_send(struct ieee80211_sub_if_data *sdata)
 {
        struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
        const u8 *ssid;
+        u8 *dst = ifmgd->associated->bssid;
+        u8 unicast_limit = max(1, IEEE80211_MAX_PROBE_TRIES - 3);
+        /*
+         * Try sending broadcast probe requests for the last three
+         * probe requests after the first ones failed since some
+         * buggy APs only support broadcast probe requests.
+         */
+        if (ifmgd->probe_send_count >= unicast_limit)
+                dst = NULL;
        ssid = ieee80211_bss_get_ie(ifmgd->associated, WLAN_EID_SSID);
-        ieee80211_send_probe_req(sdata, ifmgd->associated->bssid,
+        ieee80211_send_probe_req(sdata, dst, ssid + 2, ssid[1], NULL, 0);
-                                 ssid + 2, ssid[1], NULL, 0);
        ifmgd->probe_send_count++;
        ifmgd->probe_timeout = jiffies + IEEE80211_PROBE_WAIT;
@@ -1103,8 +1125,11 @@ static void __ieee80211_connection_loss(struct ieee80211_sub_if_data *sdata)
        printk(KERN_DEBUG "Connection to AP %pM lost.\n", bssid);
        ieee80211_set_disassoc(sdata, true);
-        ieee80211_recalc_idle(local);
        mutex_unlock(&ifmgd->mtx);
+        mutex_lock(&local->mtx);
+        ieee80211_recalc_idle(local);
+        mutex_unlock(&local->mtx);
        /*
         * must be outside lock due to cfg80211,
         * but that's not a problem.
@@ -1173,7 +1198,9 @@ ieee80211_rx_mgmt_deauth(struct ieee80211_sub_if_data *sdata,
                        sdata->name, bssid, reason_code);
        ieee80211_set_disassoc(sdata, true);
+        mutex_lock(&sdata->local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&sdata->local->mtx);
        return RX_MGMT_CFG80211_DEAUTH;
 }
@@ -1203,7 +1230,9 @@ ieee80211_rx_mgmt_disassoc(struct ieee80211_sub_if_data *sdata,
                        sdata->name, mgmt->sa, reason_code);
        ieee80211_set_disassoc(sdata, true);
+        mutex_lock(&sdata->local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&sdata->local->mtx);
        return RX_MGMT_CFG80211_DISASSOC;
 }
@@ -1330,6 +1359,14 @@ static bool ieee80211_assoc_success(struct ieee80211_work *wk,
                return false;
        }
+        /*
+         * Always handle WMM once after association regardless
+         * of the first value the AP uses. Setting -1 here has
+         * that effect because the AP values is an unsigned
+         * 4-bit value.
+         */
+        ifmgd->wmm_last_param_set = -1;
        if (elems.wmm_param)
                ieee80211_sta_wmm_params(local, sdata, elems.wmm_param,
                                         elems.wmm_param_len);
@@ -1362,7 +1399,7 @@ static bool ieee80211_assoc_success(struct ieee80211_work *wk,
         * Also start the timer that will detect beacon loss.
         */
        ieee80211_sta_rx_notify(sdata, (struct ieee80211_hdr *)mgmt);
-        mod_beacon_timer(sdata);
+        ieee80211_sta_reset_beacon_monitor(sdata);
        return true;
 }
@@ -1465,7 +1502,7 @@ static void ieee80211_rx_mgmt_probe_resp(struct ieee80211_sub_if_data *sdata,
                 * we have or will be receiving any beacons or data, so let's
                 * schedule the timers again, just in case.
                 */
-                mod_beacon_timer(sdata);
+                ieee80211_sta_reset_beacon_monitor(sdata);
                mod_timer(&ifmgd->conn_mon_timer,
                          round_jiffies_up(jiffies +
@@ -1540,15 +1577,18 @@ static void ieee80211_rx_mgmt_beacon(struct ieee80211_sub_if_data *sdata,
        ifmgd->last_beacon_signal = rx_status->signal;
        if (ifmgd->flags & IEEE80211_STA_RESET_SIGNAL_AVE) {
                ifmgd->flags &= ~IEEE80211_STA_RESET_SIGNAL_AVE;
-                ifmgd->ave_beacon_signal = rx_status->signal;
+                ifmgd->ave_beacon_signal = rx_status->signal * 16;
                ifmgd->last_cqm_event_signal = 0;
+                ifmgd->count_beacon_signal = 1;
        } else {
                ifmgd->ave_beacon_signal =
                        (IEEE80211_SIGNAL_AVE_WEIGHT * rx_status->signal * 16 +
                         (16 - IEEE80211_SIGNAL_AVE_WEIGHT) *
                         ifmgd->ave_beacon_signal) / 16;
+                ifmgd->count_beacon_signal++;
        }
        if (bss_conf->cqm_rssi_thold &&
+            ifmgd->count_beacon_signal >= IEEE80211_SIGNAL_AVE_MIN_COUNT &&
            !(local->hw.flags & IEEE80211_HW_SUPPORTS_CQM_RSSI)) {
                int sig = ifmgd->ave_beacon_signal / 16;
                int last_event = ifmgd->last_cqm_event_signal;
@@ -1588,7 +1628,7 @@ static void ieee80211_rx_mgmt_beacon(struct ieee80211_sub_if_data *sdata,
         * Push the beacon loss detection into the future since
         * we are processing a beacon from the AP just now.
         */
-        mod_beacon_timer(sdata);
+        ieee80211_sta_reset_beacon_monitor(sdata);
        ncrc = crc32_be(0, (void *)&mgmt->u.beacon.beacon_int, 4);
        ncrc = ieee802_11_parse_elems_crc(mgmt->u.beacon.variable,
@@ -1751,7 +1791,7 @@ void ieee80211_sta_rx_queued_mgmt(struct ieee80211_sub_if_data *sdata,
                struct ieee80211_local *local = sdata->local;
                struct ieee80211_work *wk;
-                mutex_lock(&local->work_mtx);
+                mutex_lock(&local->mtx);
                list_for_each_entry(wk, &local->work_list, list) {
                        if (wk->sdata != sdata)
                                continue;
@@ -1783,7 +1823,7 @@ void ieee80211_sta_rx_queued_mgmt(struct ieee80211_sub_if_data *sdata,
                        free_work(wk);
                        break;
                }
-                mutex_unlock(&local->work_mtx);
+                mutex_unlock(&local->mtx);
                cfg80211_send_deauth(sdata->dev, (u8 *)mgmt, skb->len);
        }
@@ -1840,8 +1880,10 @@ void ieee80211_sta_work(struct ieee80211_sub_if_data *sdata)
                                " after %dms, disconnecting.\n",
                                bssid, (1000 * IEEE80211_PROBE_WAIT)/HZ);
                        ieee80211_set_disassoc(sdata, true);
-                        ieee80211_recalc_idle(local);
                        mutex_unlock(&ifmgd->mtx);
+                        mutex_lock(&local->mtx);
+                        ieee80211_recalc_idle(local);
+                        mutex_unlock(&local->mtx);
                        /*
                         * must be outside lock due to cfg80211,
                         * but that's not a problem.
@@ -1917,6 +1959,8 @@ void ieee80211_sta_quiesce(struct ieee80211_sub_if_data *sdata)
         * time -- the code here is properly synchronised.
         */
+        cancel_work_sync(&ifmgd->request_smps_work);
        cancel_work_sync(&ifmgd->beacon_connection_loss_work);
        if (del_timer_sync(&ifmgd->timer))
                set_bit(TMR_RUNNING_TIMER, &ifmgd->timers_running);
@@ -1952,6 +1996,7 @@ void ieee80211_sta_setup_sdata(struct ieee80211_sub_if_data *sdata)
        INIT_WORK(&ifmgd->chswitch_work, ieee80211_chswitch_work);
        INIT_WORK(&ifmgd->beacon_connection_loss_work,
                  ieee80211_beacon_connection_loss_work);
+        INIT_WORK(&ifmgd->request_smps_work, ieee80211_request_smps_work);
        setup_timer(&ifmgd->timer, ieee80211_sta_timer,
                    (unsigned long) sdata);
        setup_timer(&ifmgd->bcn_mon_timer, ieee80211_sta_bcn_mon_timer,
@@ -2249,6 +2294,9 @@ int ieee80211_mgd_assoc(struct ieee80211_sub_if_data *sdata,
        else
                ifmgd->flags &= ~IEEE80211_STA_CONTROL_PORT;
+        sdata->control_port_protocol = req->crypto.control_port_ethertype;
+        sdata->control_port_no_encrypt = req->crypto.control_port_no_encrypt;
        ieee80211_add_work(wk);
        return 0;
 }
@@ -2275,7 +2323,7 @@ int ieee80211_mgd_deauth(struct ieee80211_sub_if_data *sdata,
                mutex_unlock(&ifmgd->mtx);
-                mutex_lock(&local->work_mtx);
+                mutex_lock(&local->mtx);
                list_for_each_entry(wk, &local->work_list, list) {
                        if (wk->sdata != sdata)
                                continue;
@@ -2294,7 +2342,7 @@ int ieee80211_mgd_deauth(struct ieee80211_sub_if_data *sdata,
                        free_work(wk);
                        break;
                }
-                mutex_unlock(&local->work_mtx);
+                mutex_unlock(&local->mtx);
                /*
                 * If somebody requests authentication and we haven't
@@ -2319,7 +2367,9 @@ int ieee80211_mgd_deauth(struct ieee80211_sub_if_data *sdata,
        if (assoc_bss)
                sta_info_destroy_addr(sdata, bssid);
+        mutex_lock(&sdata->local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&sdata->local->mtx);
        return 0;
 }
@@ -2357,7 +2407,9 @@ int ieee80211_mgd_disassoc(struct ieee80211_sub_if_data *sdata,
                        cookie, !req->local_state_change);
        sta_info_destroy_addr(sdata, bssid);
+        mutex_lock(&sdata->local->mtx);
        ieee80211_recalc_idle(sdata->local);
+        mutex_unlock(&sdata->local->mtx);
        return 0;
 }
diff --git a/net/mac80211/offchannel.c b/net/mac80211/offchannel.c
index c36b1911987a..4b564091e51d 100644
--- a/net/mac80211/offchannel.c
+++ b/net/mac80211/offchannel.c
@@ -22,12 +22,16 @@
 static void ieee80211_offchannel_ps_enable(struct ieee80211_sub_if_data *sdata)
 {
        struct ieee80211_local *local = sdata->local;
+        struct ieee80211_if_managed *ifmgd = &sdata->u.mgd;
        local->offchannel_ps_enabled = false;
        /* FIXME: what to do when local->pspolling is true? */
        del_timer_sync(&local->dynamic_ps_timer);
+        del_timer_sync(&ifmgd->bcn_mon_timer);
+        del_timer_sync(&ifmgd->conn_mon_timer);
        cancel_work_sync(&local->dynamic_ps_enable_work);
        if (local->hw.conf.flags & IEEE80211_CONF_PS) {
@@ -85,6 +89,9 @@ static void ieee80211_offchannel_ps_disable(struct ieee80211_sub_if_data *sdata)
                mod_timer(&local->dynamic_ps_timer, jiffies +
                          msecs_to_jiffies(local->hw.conf.dynamic_ps_timeout));
        }
+        ieee80211_sta_reset_beacon_monitor(sdata);
+        ieee80211_sta_reset_conn_monitor(sdata);
 }
 void ieee80211_offchannel_stop_beaconing(struct ieee80211_local *local)
@@ -112,8 +119,10 @@ void ieee80211_offchannel_stop_beaconing(struct ieee80211_local *local)
                 * used from user space controlled off-channel operations.
                 */
                if (sdata->vif.type != NL80211_IFTYPE_STATION &&
-                    sdata->vif.type != NL80211_IFTYPE_MONITOR)
+                    sdata->vif.type != NL80211_IFTYPE_MONITOR) {
+                        set_bit(SDATA_STATE_OFFCHANNEL, &sdata->state);
                        netif_tx_stop_all_queues(sdata->dev);
+                }
        }
        mutex_unlock(&local->iflist_mtx);
 }
@@ -131,6 +140,7 @@ void ieee80211_offchannel_stop_station(struct ieee80211_local *local)
                        continue;
                if (sdata->vif.type == NL80211_IFTYPE_STATION) {
+                        set_bit(SDATA_STATE_OFFCHANNEL, &sdata->state);
                        netif_tx_stop_all_queues(sdata->dev);
                        if (sdata->u.mgd.associated)
                                ieee80211_offchannel_ps_enable(sdata);
@@ -155,8 +165,20 @@ void ieee80211_offchannel_return(struct ieee80211_local *local,
                                ieee80211_offchannel_ps_disable(sdata);
                }
-                if (sdata->vif.type != NL80211_IFTYPE_MONITOR)
+                if (sdata->vif.type != NL80211_IFTYPE_MONITOR) {
+                        clear_bit(SDATA_STATE_OFFCHANNEL, &sdata->state);
+                        /*
+                         * This may wake up queues even though the driver
+                         * currently has them stopped. This is not very
+                         * likely, since the driver won't have gotten any
+                         * (or hardly any) new packets while we weren't
+                         * on the right channel, and even if it happens
+                         * it will at most lead to queueing up one more
+                         * packet per queue in mac80211 rather than on
+                         * the interface qdisc.
+                         */
                        netif_tx_wake_all_queues(sdata->dev);
+                }
                /* re-enable beaconing */
                if (enable_beaconing &&
diff --git a/net/mac80211/pm.c b/net/mac80211/pm.c
index d287fde0431d..ce671dfd238c 100644
--- a/net/mac80211/pm.c
+++ b/net/mac80211/pm.c
@@ -12,7 +12,8 @@ int __ieee80211_suspend(struct ieee80211_hw *hw)
        struct ieee80211_sub_if_data *sdata;
        struct sta_info *sta;
-        ieee80211_scan_cancel(local);
+        if (unlikely(test_bit(SCAN_SW_SCANNING, &local->scanning)))
+                ieee80211_scan_cancel(local);
        ieee80211_stop_queues_by_reason(hw,
                        IEEE80211_QUEUE_STOP_REASON_SUSPEND);
diff --git a/net/mac80211/rate.c b/net/mac80211/rate.c
index be04d46110fe..b0cc385bf989 100644
--- a/net/mac80211/rate.c
+++ b/net/mac80211/rate.c
@@ -207,7 +207,7 @@ static bool rc_no_data_or_no_ack(struct ieee80211_tx_rate_control *txrc)
        fc = hdr->frame_control;
-        return ((info->flags & IEEE80211_TX_CTL_NO_ACK) || !ieee80211_is_data(fc));
+        return (info->flags & IEEE80211_TX_CTL_NO_ACK) || !ieee80211_is_data(fc);
 }
 static void rc_send_low_broadcast(s8 *idx, u32 basic_rates, u8 max_rate_idx)
@@ -368,8 +368,8 @@ int ieee80211_init_rate_ctrl_alg(struct ieee80211_local *local,
        ref = rate_control_alloc(name, local);
        if (!ref) {
-                printk(KERN_WARNING "%s: Failed to select rate control "
+                wiphy_warn(local->hw.wiphy,
-                       "algorithm\n", wiphy_name(local->hw.wiphy));
+                           "Failed to select rate control algorithm\n");
                return -ENOENT;
        }
@@ -380,9 +380,8 @@ int ieee80211_init_rate_ctrl_alg(struct ieee80211_local *local,
                sta_info_flush(local, NULL);
        }
-        printk(KERN_DEBUG "%s: Selected rate control "
+        wiphy_debug(local->hw.wiphy, "Selected rate control algorithm '%s'\n",
-               "algorithm '%s'\n", wiphy_name(local->hw.wiphy),
+                    ref->ops->name);
-               ref->ops->name);
        return 0;
 }
diff --git a/net/mac80211/rc80211_pid_debugfs.c b/net/mac80211/rc80211_pid_debugfs.c
index 47438b4a9af5..135f36fd4d5d 100644
--- a/net/mac80211/rc80211_pid_debugfs.c
+++ b/net/mac80211/rc80211_pid_debugfs.c
@@ -162,7 +162,7 @@ static ssize_t rate_control_pid_events_read(struct file *file, char __user *buf,
        file_info->next_entry = (file_info->next_entry + 1) %
                                RC_PID_EVENT_RING_SIZE;
-        /* Print information about the event. Note that userpace needs to
+        /* Print information about the event. Note that userspace needs to
         * provide large enough buffers. */
        length = length < RC_PID_PRINT_BUF_SIZE ?
                 length : RC_PID_PRINT_BUF_SIZE;
diff --git a/net/mac80211/rx.c b/net/mac80211/rx.c
index 28624282c5f3..0b0e83ebe3d5 100644
--- a/net/mac80211/rx.c
+++ b/net/mac80211/rx.c
@@ -315,6 +315,7 @@ ieee80211_rx_monitor(struct ieee80211_local *local, struct sk_buff *origskb,
 static void ieee80211_parse_qos(struct ieee80211_rx_data *rx)
 {
        struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        int tid;
        /* does the frame have a qos control field? */
@@ -323,9 +324,7 @@ static void ieee80211_parse_qos(struct ieee80211_rx_data *rx)
                /* frame has qos control */
                tid = *qc & IEEE80211_QOS_CTL_TID_MASK;
                if (*qc & IEEE80211_QOS_CONTROL_A_MSDU_PRESENT)
-                        rx->flags |= IEEE80211_RX_AMSDU;
+                        status->rx_flags |= IEEE80211_RX_AMSDU;
-                else
-                        rx->flags &= ~IEEE80211_RX_AMSDU;
        } else {
                /*
                 * IEEE 802.11-2007, 7.1.3.4.1 ("Sequence Number field"):
@@ -387,26 +386,25 @@ static ieee80211_rx_result debug_noinline
 ieee80211_rx_h_passive_scan(struct ieee80211_rx_data *rx)
 {
        struct ieee80211_local *local = rx->local;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        struct sk_buff *skb = rx->skb;
-        if (unlikely(test_bit(SCAN_HW_SCANNING, &local->scanning)))
+        if (likely(!(status->rx_flags & IEEE80211_RX_IN_SCAN)))
+                return RX_CONTINUE;
+        if (test_bit(SCAN_HW_SCANNING, &local->scanning))
                return ieee80211_scan_rx(rx->sdata, skb);
-        if (unlikely(test_bit(SCAN_SW_SCANNING, &local->scanning) &&
+        if (test_bit(SCAN_SW_SCANNING, &local->scanning)) {
-                     (rx->flags & IEEE80211_RX_IN_SCAN))) {
                /* drop all the other packets during a software scan anyway */
                if (ieee80211_scan_rx(rx->sdata, skb) != RX_QUEUED)
                        dev_kfree_skb(skb);
                return RX_QUEUED;
        }
-        if (unlikely(rx->flags & IEEE80211_RX_IN_SCAN)) {
+        /* scanning finished during invoking of handlers */
-                /* scanning finished during invoking of handlers */
+        I802_DEBUG_INC(local->rx_handlers_drop_passive_scan);
-                I802_DEBUG_INC(local->rx_handlers_drop_passive_scan);
+        return RX_DROP_UNUSABLE;
-                return RX_DROP_UNUSABLE;
-        }
-        return RX_CONTINUE;
 }
@@ -538,20 +536,12 @@ static void ieee80211_release_reorder_frame(struct ieee80211_hw *hw,
                                            int index,
                                            struct sk_buff_head *frames)
 {
-        struct ieee80211_supported_band *sband;
-        struct ieee80211_rate *rate = NULL;
        struct sk_buff *skb = tid_agg_rx->reorder_buf[index];
-        struct ieee80211_rx_status *status;
        if (!skb)
                goto no_frame;
-        status = IEEE80211_SKB_RXCB(skb);
+        /* release the frame from the reorder ring buffer */
-        /* release the reordered frames to stack */
-        sband = hw->wiphy->bands[status->band];
-        if (!(status->flag & RX_FLAG_HT))
-                rate = &sband->bitrates[status->rate_idx];
        tid_agg_rx->stored_mpdu_num--;
        tid_agg_rx->reorder_buf[index] = NULL;
        __skb_queue_tail(frames, skb);
@@ -580,9 +570,78 @@ static void ieee80211_release_reorder_frames(struct ieee80211_hw *hw,
 * frames that have not yet been received are assumed to be lost and the skb
 * can be released for processing. This may also release other skb's from the
 * reorder buffer if there are no additional gaps between the frames.
+ *
+ * Callers must hold tid_agg_rx->reorder_lock.
 */
 #define HT_RX_REORDER_BUF_TIMEOUT (HZ / 10)
+static void ieee80211_sta_reorder_release(struct ieee80211_hw *hw,
+                                          struct tid_ampdu_rx *tid_agg_rx,
+                                          struct sk_buff_head *frames)
+{
+        int index, j;
+        /* release the buffer until next missing frame */
+        index = seq_sub(tid_agg_rx->head_seq_num, tid_agg_rx->ssn) %
+                                                tid_agg_rx->buf_size;
+        if (!tid_agg_rx->reorder_buf[index] &&
+            tid_agg_rx->stored_mpdu_num > 1) {
+                /*
+                 * No buffers ready to be released, but check whether any
+                 * frames in the reorder buffer have timed out.
+                 */
+                int skipped = 1;
+                for (j = (index + 1) % tid_agg_rx->buf_size; j != index;
+                     j = (j + 1) % tid_agg_rx->buf_size) {
+                        if (!tid_agg_rx->reorder_buf[j]) {
+                                skipped++;
+                                continue;
+                        }
+                        if (!time_after(jiffies, tid_agg_rx->reorder_time[j] +
+                                        HT_RX_REORDER_BUF_TIMEOUT))
+                                goto set_release_timer;
+#ifdef CONFIG_MAC80211_HT_DEBUG
+                        if (net_ratelimit())
+                                wiphy_debug(hw->wiphy,
+                                            "release an RX reorder frame due to timeout on earlier frames\n");
+#endif
+                        ieee80211_release_reorder_frame(hw, tid_agg_rx,
+                                                        j, frames);
+                        /*
+                         * Increment the head seq# also for the skipped slots.
+                         */
+                        tid_agg_rx->head_seq_num =
+                                (tid_agg_rx->head_seq_num + skipped) & SEQ_MASK;
+                        skipped = 0;
+                }
+        } else while (tid_agg_rx->reorder_buf[index]) {
+                ieee80211_release_reorder_frame(hw, tid_agg_rx, index, frames);
+                index = seq_sub(tid_agg_rx->head_seq_num, tid_agg_rx->ssn) %
+                                                        tid_agg_rx->buf_size;
+        }
+        if (tid_agg_rx->stored_mpdu_num) {
+                j = index = seq_sub(tid_agg_rx->head_seq_num,
+                                    tid_agg_rx->ssn) % tid_agg_rx->buf_size;
+                for (; j != (index - 1) % tid_agg_rx->buf_size;
+                     j = (j + 1) % tid_agg_rx->buf_size) {
+                        if (tid_agg_rx->reorder_buf[j])
+                                break;
+                }
+ set_release_timer:
+                mod_timer(&tid_agg_rx->reorder_timer,
+                          tid_agg_rx->reorder_time[j] +
+                          HT_RX_REORDER_BUF_TIMEOUT);
+        } else {
+                del_timer(&tid_agg_rx->reorder_timer);
+        }
+}
 /*
 * As this function belongs to the RX path it must be under
 * rcu_read_lock protection. It returns false if the frame
@@ -598,14 +657,16 @@ static bool ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw,
        u16 mpdu_seq_num = (sc & IEEE80211_SCTL_SEQ) >> 4;
        u16 head_seq_num, buf_size;
        int index;
+        bool ret = true;
        buf_size = tid_agg_rx->buf_size;
        head_seq_num = tid_agg_rx->head_seq_num;
+        spin_lock(&tid_agg_rx->reorder_lock);
        /* frame with out of date sequence number */
        if (seq_less(mpdu_seq_num, head_seq_num)) {
                dev_kfree_skb(skb);
-                return true;
+                goto out;
        }
        /*
@@ -626,7 +687,7 @@ static bool ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw,
        /* check if we already stored this frame */
        if (tid_agg_rx->reorder_buf[index]) {
                dev_kfree_skb(skb);
-                return true;
+                goto out;
        }
        /*
@@ -636,58 +697,19 @@ static bool ieee80211_sta_manage_reorder_buf(struct ieee80211_hw *hw,
        if (mpdu_seq_num == tid_agg_rx->head_seq_num &&
            tid_agg_rx->stored_mpdu_num == 0) {
                tid_agg_rx->head_seq_num = seq_inc(tid_agg_rx->head_seq_num);
-                return false;
+                ret = false;
+                goto out;
        }
        /* put the frame in the reordering buffer */
        tid_agg_rx->reorder_buf[index] = skb;
        tid_agg_rx->reorder_time[index] = jiffies;
        tid_agg_rx->stored_mpdu_num++;
-        /* release the buffer until next missing frame */
+        ieee80211_sta_reorder_release(hw, tid_agg_rx, frames);
-        index = seq_sub(tid_agg_rx->head_seq_num, tid_agg_rx->ssn) %
-                                                tid_agg_rx->buf_size;
-        if (!tid_agg_rx->reorder_buf[index] &&
-            tid_agg_rx->stored_mpdu_num > 1) {
-                /*
-                 * No buffers ready to be released, but check whether any
-                 * frames in the reorder buffer have timed out.
-                 */
-                int j;
-                int skipped = 1;
-                for (j = (index + 1) % tid_agg_rx->buf_size; j != index;
-                     j = (j + 1) % tid_agg_rx->buf_size) {
-                        if (!tid_agg_rx->reorder_buf[j]) {
-                                skipped++;
-                                continue;
-                        }
-                        if (!time_after(jiffies, tid_agg_rx->reorder_time[j] +
-                                        HT_RX_REORDER_BUF_TIMEOUT))
-                                break;
-#ifdef CONFIG_MAC80211_HT_DEBUG
-                        if (net_ratelimit())
-                                printk(KERN_DEBUG "%s: release an RX reorder "
-                                       "frame due to timeout on earlier "
-                                       "frames\n",
-                                       wiphy_name(hw->wiphy));
-#endif
-                        ieee80211_release_reorder_frame(hw, tid_agg_rx,
-                                                        j, frames);
-                        /*
+ out:
-                         * Increment the head seq# also for the skipped slots.
+        spin_unlock(&tid_agg_rx->reorder_lock);
-                         */
+        return ret;
-                        tid_agg_rx->head_seq_num =
-                                (tid_agg_rx->head_seq_num + skipped) & SEQ_MASK;
-                        skipped = 0;
-                }
-        } else while (tid_agg_rx->reorder_buf[index]) {
-                ieee80211_release_reorder_frame(hw, tid_agg_rx, index, frames);
-                index = seq_sub(tid_agg_rx->head_seq_num, tid_agg_rx->ssn) %
-                                                        tid_agg_rx->buf_size;
-        }
-        return true;
 }
 /*
@@ -761,13 +783,14 @@ static ieee80211_rx_result debug_noinline
 ieee80211_rx_h_check(struct ieee80211_rx_data *rx)
 {
        struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)rx->skb->data;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        /* Drop duplicate 802.11 retransmissions (IEEE 802.11 Chap. 9.2.9) */
        if (rx->sta && !is_multicast_ether_addr(hdr->addr1)) {
                if (unlikely(ieee80211_has_retry(hdr->frame_control) &&
                             rx->sta->last_seq_ctrl[rx->queue] ==
                             hdr->seq_ctrl)) {
-                        if (rx->flags & IEEE80211_RX_RA_MATCH) {
+                        if (status->rx_flags & IEEE80211_RX_RA_MATCH) {
                                rx->local->dot11FrameDuplicateCount++;
                                rx->sta->num_duplicates++;
                        }
@@ -800,7 +823,7 @@ ieee80211_rx_h_check(struct ieee80211_rx_data *rx)
                if ((!ieee80211_has_fromds(hdr->frame_control) &&
                     !ieee80211_has_tods(hdr->frame_control) &&
                     ieee80211_is_data(hdr->frame_control)) ||
-                    !(rx->flags & IEEE80211_RX_RA_MATCH)) {
+                    !(status->rx_flags & IEEE80211_RX_RA_MATCH)) {
                        /* Drop IBSS frames and frames for other hosts
                         * silently. */
                        return RX_DROP_MONITOR;
@@ -857,7 +880,7 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx)
         * No point in finding a key and decrypting if the frame is neither
         * addressed to us nor a multicast frame.
         */
-        if (!(rx->flags & IEEE80211_RX_RA_MATCH))
+        if (!(status->rx_flags & IEEE80211_RX_RA_MATCH))
                return RX_CONTINUE;
        /* start without a key */
@@ -873,6 +896,9 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx)
        if (!is_multicast_ether_addr(hdr->addr1) && stakey) {
                rx->key = stakey;
+                if ((status->flag & RX_FLAG_DECRYPTED) &&
+                    (status->flag & RX_FLAG_IV_STRIPPED))
+                        return RX_CONTINUE;
                /* Skip decryption if the frame is not protected. */
                if (!ieee80211_has_protected(fc))
                        return RX_CONTINUE;
@@ -935,7 +961,8 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx)
                 * pairwise or station-to-station keys, but for WEP we allow
                 * using a key index as well.
                 */
-                if (rx->key && rx->key->conf.alg != ALG_WEP &&
+                if (rx->key && rx->key->conf.cipher != WLAN_CIPHER_SUITE_WEP40 &&
+                    rx->key->conf.cipher != WLAN_CIPHER_SUITE_WEP104 &&
                    !is_multicast_ether_addr(hdr->addr1))
                        rx->key = NULL;
        }
@@ -951,8 +978,9 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx)
                return RX_DROP_UNUSABLE;
        /* the hdr variable is invalid now! */
-        switch (rx->key->conf.alg) {
+        switch (rx->key->conf.cipher) {
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_WEP40:
+        case WLAN_CIPHER_SUITE_WEP104:
                /* Check for weak IVs if possible */
                if (rx->sta && ieee80211_is_data(fc) &&
                    (!(status->flag & RX_FLAG_IV_STRIPPED) ||
@@ -962,15 +990,21 @@ ieee80211_rx_h_decrypt(struct ieee80211_rx_data *rx)
                result = ieee80211_crypto_wep_decrypt(rx);
                break;
-        case ALG_TKIP:
+        case WLAN_CIPHER_SUITE_TKIP:
                result = ieee80211_crypto_tkip_decrypt(rx);
                break;
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                result = ieee80211_crypto_ccmp_decrypt(rx);
                break;
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                result = ieee80211_crypto_aes_cmac_decrypt(rx);
                break;
+        default:
+                /*
+                 * We can reach here only with HW-only algorithms
+                 * but why didn't it decrypt the frame?!
+                 */
+                return RX_DROP_UNUSABLE;
        }
        /* either the frame has been decrypted or will be dropped */
@@ -1079,7 +1113,7 @@ ieee80211_rx_h_sta_process(struct ieee80211_rx_data *rx)
                sta->last_rx = jiffies;
        }
-        if (!(rx->flags & IEEE80211_RX_RA_MATCH))
+        if (!(status->rx_flags & IEEE80211_RX_RA_MATCH))
                return RX_CONTINUE;
        if (rx->sdata->vif.type == NL80211_IFTYPE_STATION)
@@ -1236,6 +1270,7 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx)
        unsigned int frag, seq;
        struct ieee80211_fragment_entry *entry;
        struct sk_buff *skb;
+        struct ieee80211_rx_status *status;
        hdr = (struct ieee80211_hdr *)rx->skb->data;
        fc = hdr->frame_control;
@@ -1265,7 +1300,7 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx)
                /* This is the first fragment of a new frame. */
                entry = ieee80211_reassemble_add(rx->sdata, frag, seq,
                                                 rx->queue, &(rx->skb));
-                if (rx->key && rx->key->conf.alg == ALG_CCMP &&
+                if (rx->key && rx->key->conf.cipher == WLAN_CIPHER_SUITE_CCMP &&
                    ieee80211_has_protected(fc)) {
                        int queue = ieee80211_is_mgmt(fc) ?
                                NUM_RX_DATA_QUEUES : rx->queue;
@@ -1294,7 +1329,7 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx)
                int i;
                u8 pn[CCMP_PN_LEN], *rpn;
                int queue;
-                if (!rx->key || rx->key->conf.alg != ALG_CCMP)
+                if (!rx->key || rx->key->conf.cipher != WLAN_CIPHER_SUITE_CCMP)
                        return RX_DROP_UNUSABLE;
                memcpy(pn, entry->last_pn, CCMP_PN_LEN);
                for (i = CCMP_PN_LEN - 1; i >= 0; i--) {
@@ -1335,7 +1370,8 @@ ieee80211_rx_h_defragment(struct ieee80211_rx_data *rx)
        }
        /* Complete frame has been reassembled - process it now */
-        rx->flags |= IEEE80211_RX_FRAGMENTED;
+        status = IEEE80211_SKB_RXCB(rx->skb);
+        status->rx_flags |= IEEE80211_RX_FRAGMENTED;
 out:
        if (rx->sta)
@@ -1352,9 +1388,10 @@ ieee80211_rx_h_ps_poll(struct ieee80211_rx_data *rx)
 {
        struct ieee80211_sub_if_data *sdata = rx->sdata;
        __le16 fc = ((struct ieee80211_hdr *)rx->skb->data)->frame_control;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        if (likely(!rx->sta || !ieee80211_is_pspoll(fc) ||
-                   !(rx->flags & IEEE80211_RX_RA_MATCH)))
+                   !(status->rx_flags & IEEE80211_RX_RA_MATCH)))
                return RX_CONTINUE;
        if ((sdata->vif.type != NL80211_IFTYPE_AP) &&
@@ -1492,7 +1529,7 @@ static bool ieee80211_frame_allowed(struct ieee80211_rx_data *rx, __le16 fc)
         * Allow EAPOL frames to us/the PAE group address regardless
         * of whether the frame was encrypted or not.
         */
-        if (ehdr->h_proto == htons(ETH_P_PAE) &&
+        if (ehdr->h_proto == rx->sdata->control_port_protocol &&
            (compare_ether_addr(ehdr->h_dest, rx->sdata->vif.addr) == 0 ||
             compare_ether_addr(ehdr->h_dest, pae_group_addr) == 0))
                return true;
@@ -1515,6 +1552,7 @@ ieee80211_deliver_skb(struct ieee80211_rx_data *rx)
        struct sk_buff *skb, *xmit_skb;
        struct ethhdr *ehdr = (struct ethhdr *) rx->skb->data;
        struct sta_info *dsta;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        skb = rx->skb;
        xmit_skb = NULL;
@@ -1522,7 +1560,7 @@ ieee80211_deliver_skb(struct ieee80211_rx_data *rx)
        if ((sdata->vif.type == NL80211_IFTYPE_AP ||
             sdata->vif.type == NL80211_IFTYPE_AP_VLAN) &&
            !(sdata->flags & IEEE80211_SDATA_DONT_BRIDGE_PACKETS) &&
-            (rx->flags & IEEE80211_RX_RA_MATCH) &&
+            (status->rx_flags & IEEE80211_RX_RA_MATCH) &&
            (sdata->vif.type != NL80211_IFTYPE_AP_VLAN || !sdata->u.vlan.sta)) {
                if (is_multicast_ether_addr(ehdr->h_dest)) {
                        /*
@@ -1599,6 +1637,7 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx)
        struct ieee80211_hdr *hdr = (struct ieee80211_hdr *)skb->data;
        __le16 fc = hdr->frame_control;
        struct sk_buff_head frame_list;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
        if (unlikely(!ieee80211_is_data(fc)))
                return RX_CONTINUE;
@@ -1606,7 +1645,7 @@ ieee80211_rx_h_amsdu(struct ieee80211_rx_data *rx)
        if (unlikely(!ieee80211_is_data_present(fc)))
                return RX_DROP_MONITOR;
-        if (!(rx->flags & IEEE80211_RX_AMSDU))
+        if (!(status->rx_flags & IEEE80211_RX_AMSDU))
                return RX_CONTINUE;
        if (ieee80211_has_a4(hdr->frame_control) &&
@@ -1657,6 +1696,7 @@ ieee80211_rx_h_mesh_fwding(struct ieee80211_rx_data *rx)
        struct sk_buff *skb = rx->skb, *fwd_skb;
        struct ieee80211_local *local = rx->local;
        struct ieee80211_sub_if_data *sdata = rx->sdata;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
        hdr = (struct ieee80211_hdr *) skb->data;
        hdrlen = ieee80211_hdrlen(hdr->frame_control);
@@ -1702,7 +1742,7 @@ ieee80211_rx_h_mesh_fwding(struct ieee80211_rx_data *rx)
        mesh_hdr->ttl--;
-        if (rx->flags & IEEE80211_RX_RA_MATCH) {
+        if (status->rx_flags & IEEE80211_RX_RA_MATCH) {
                if (!mesh_hdr->ttl)
                        IEEE80211_IFSTA_MESH_CTR_INC(&rx->sdata->u.mesh,
                                                     dropped_frames_ttl);
@@ -1909,13 +1949,38 @@ static void ieee80211_process_sa_query_req(struct ieee80211_sub_if_data *sdata,
 }
 static ieee80211_rx_result debug_noinline
+ieee80211_rx_h_mgmt_check(struct ieee80211_rx_data *rx)
+{
+        struct ieee80211_mgmt *mgmt = (struct ieee80211_mgmt *) rx->skb->data;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
+        /*
+         * From here on, look only at management frames.
+         * Data and control frames are already handled,
+         * and unknown (reserved) frames are useless.
+         */
+        if (rx->skb->len < 24)
+                return RX_DROP_MONITOR;
+        if (!ieee80211_is_mgmt(mgmt->frame_control))
+                return RX_DROP_MONITOR;
+        if (!(status->rx_flags & IEEE80211_RX_RA_MATCH))
+                return RX_DROP_MONITOR;
+        if (ieee80211_drop_unencrypted_mgmt(rx))
+                return RX_DROP_UNUSABLE;
+        return RX_CONTINUE;
+}
+static ieee80211_rx_result debug_noinline
 ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
 {
        struct ieee80211_local *local = rx->local;
        struct ieee80211_sub_if_data *sdata = rx->sdata;
        struct ieee80211_mgmt *mgmt = (struct ieee80211_mgmt *) rx->skb->data;
-        struct sk_buff *nskb;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
-        struct ieee80211_rx_status *status;
        int len = rx->skb->len;
        if (!ieee80211_is_action(mgmt->frame_control))
@@ -1928,10 +1993,7 @@ ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
        if (!rx->sta && mgmt->u.action.category != WLAN_CATEGORY_PUBLIC)
                return RX_DROP_UNUSABLE;
-        if (!(rx->flags & IEEE80211_RX_RA_MATCH))
+        if (!(status->rx_flags & IEEE80211_RX_RA_MATCH))
-                return RX_DROP_UNUSABLE;
-        if (ieee80211_drop_unencrypted_mgmt(rx))
                return RX_DROP_UNUSABLE;
        switch (mgmt->u.action.category) {
@@ -2024,17 +2086,36 @@ ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
                goto queue;
        }
+        return RX_CONTINUE;
 invalid:
-        /*
+        status->rx_flags |= IEEE80211_RX_MALFORMED_ACTION_FRM;
-         * For AP mode, hostapd is responsible for handling any action
+        /* will return in the next handlers */
-         * frames that we didn't handle, including returning unknown
+        return RX_CONTINUE;
-         * ones. For all other modes we will return them to the sender,
-         * setting the 0x80 bit in the action category, as required by
+ handled:
-         * 802.11-2007 7.3.1.11.
+        if (rx->sta)
-         */
+                rx->sta->rx_packets++;
-        if (sdata->vif.type == NL80211_IFTYPE_AP ||
+        dev_kfree_skb(rx->skb);
-            sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
+        return RX_QUEUED;
-                return RX_DROP_MONITOR;
+ queue:
+        rx->skb->pkt_type = IEEE80211_SDATA_QUEUE_TYPE_FRAME;
+        skb_queue_tail(&sdata->skb_queue, rx->skb);
+        ieee80211_queue_work(&local->hw, &sdata->work);
+        if (rx->sta)
+                rx->sta->rx_packets++;
+        return RX_QUEUED;
+}
+static ieee80211_rx_result debug_noinline
+ieee80211_rx_h_userspace_mgmt(struct ieee80211_rx_data *rx)
+{
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
+        /* skip known-bad action frames and return them in the next handler */
+        if (status->rx_flags & IEEE80211_RX_MALFORMED_ACTION_FRM)
+                return RX_CONTINUE;
        /*
         * Getting here means the kernel doesn't know how to handle
@@ -2042,12 +2123,46 @@ ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
         * so userspace can register for those to know whether ones
         * it transmitted were processed or returned.
         */
-        status = IEEE80211_SKB_RXCB(rx->skb);
-        if (cfg80211_rx_action(rx->sdata->dev, status->freq,
+        if (cfg80211_rx_mgmt(rx->sdata->dev, status->freq,
-                               rx->skb->data, rx->skb->len,
+                             rx->skb->data, rx->skb->len,
-                               GFP_ATOMIC))
+                             GFP_ATOMIC)) {
-                goto handled;
+                if (rx->sta)
+                        rx->sta->rx_packets++;
+                dev_kfree_skb(rx->skb);
+                return RX_QUEUED;
+        }
+        return RX_CONTINUE;
+}
+static ieee80211_rx_result debug_noinline
+ieee80211_rx_h_action_return(struct ieee80211_rx_data *rx)
+{
+        struct ieee80211_local *local = rx->local;
+        struct ieee80211_mgmt *mgmt = (struct ieee80211_mgmt *) rx->skb->data;
+        struct sk_buff *nskb;
+        struct ieee80211_sub_if_data *sdata = rx->sdata;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(rx->skb);
+        if (!ieee80211_is_action(mgmt->frame_control))
+                return RX_CONTINUE;
+        /*
+         * For AP mode, hostapd is responsible for handling any action
+         * frames that we didn't handle, including returning unknown
+         * ones. For all other modes we will return them to the sender,
+         * setting the 0x80 bit in the action category, as required by
+         * 802.11-2007 7.3.1.11.
+         * Newer versions of hostapd shall also use the management frame
+         * registration mechanisms, but older ones still use cooked
+         * monitor interfaces so push all frames there.
+         */
+        if (!(status->rx_flags & IEEE80211_RX_MALFORMED_ACTION_FRM) &&
+            (sdata->vif.type == NL80211_IFTYPE_AP ||
+             sdata->vif.type == NL80211_IFTYPE_AP_VLAN))
+                return RX_DROP_MONITOR;
        /* do not return rejected action frames */
        if (mgmt->u.action.category & 0x80)
@@ -2066,20 +2181,8 @@ ieee80211_rx_h_action(struct ieee80211_rx_data *rx)
                ieee80211_tx_skb(rx->sdata, nskb);
        }
- handled:
-        if (rx->sta)
-                rx->sta->rx_packets++;
        dev_kfree_skb(rx->skb);
        return RX_QUEUED;
- queue:
-        rx->skb->pkt_type = IEEE80211_SDATA_QUEUE_TYPE_FRAME;
-        skb_queue_tail(&sdata->skb_queue, rx->skb);
-        ieee80211_queue_work(&local->hw, &sdata->work);
-        if (rx->sta)
-                rx->sta->rx_packets++;
-        return RX_QUEUED;
 }
 static ieee80211_rx_result debug_noinline
@@ -2090,15 +2193,6 @@ ieee80211_rx_h_mgmt(struct ieee80211_rx_data *rx)
        struct ieee80211_mgmt *mgmt = (void *)rx->skb->data;
        __le16 stype;
-        if (!(rx->flags & IEEE80211_RX_RA_MATCH))
-                return RX_DROP_MONITOR;
-        if (rx->skb->len < 24)
-                return RX_DROP_MONITOR;
-        if (ieee80211_drop_unencrypted_mgmt(rx))
-                return RX_DROP_UNUSABLE;
        rxs = ieee80211_work_rx_mgmt(rx->sdata, rx->skb);
        if (rxs != RX_CONTINUE)
                return rxs;
@@ -2199,6 +2293,14 @@ static void ieee80211_rx_cooked_monitor(struct ieee80211_rx_data *rx,
        struct net_device *prev_dev = NULL;
        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
+        /*
+         * If cooked monitor has been processed already, then
+         * don't do it again. If not, set the flag.
+         */
+        if (rx->flags & IEEE80211_RX_CMNTR)
+                goto out_free_skb;
+        rx->flags |= IEEE80211_RX_CMNTR;
        if (skb_headroom(skb) < sizeof(*rthdr) &&
            pskb_expand_head(skb, sizeof(*rthdr), 0, GFP_ATOMIC))
                goto out_free_skb;
@@ -2253,29 +2355,53 @@ static void ieee80211_rx_cooked_monitor(struct ieee80211_rx_data *rx,
        if (prev_dev) {
                skb->dev = prev_dev;
                netif_receive_skb(skb);
-                skb = NULL;
+                return;
-        } else
+        }
-                goto out_free_skb;
-        return;
 out_free_skb:
        dev_kfree_skb(skb);
 }
+static void ieee80211_rx_handlers_result(struct ieee80211_rx_data *rx,
+                                         ieee80211_rx_result res)
+{
+        switch (res) {
+        case RX_DROP_MONITOR:
+                I802_DEBUG_INC(rx->sdata->local->rx_handlers_drop);
+                if (rx->sta)
+                        rx->sta->rx_dropped++;
+                /* fall through */
+        case RX_CONTINUE: {
+                struct ieee80211_rate *rate = NULL;
+                struct ieee80211_supported_band *sband;
+                struct ieee80211_rx_status *status;
+                status = IEEE80211_SKB_RXCB((rx->skb));
+                sband = rx->local->hw.wiphy->bands[status->band];
+                if (!(status->flag & RX_FLAG_HT))
+                        rate = &sband->bitrates[status->rate_idx];
+                ieee80211_rx_cooked_monitor(rx, rate);
+                break;
+                }
+        case RX_DROP_UNUSABLE:
+                I802_DEBUG_INC(rx->sdata->local->rx_handlers_drop);
+                if (rx->sta)
+                        rx->sta->rx_dropped++;
+                dev_kfree_skb(rx->skb);
+                break;
+        case RX_QUEUED:
+                I802_DEBUG_INC(rx->sdata->local->rx_handlers_queued);
+                break;
+        }
+}
-static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_data *sdata,
+static void ieee80211_rx_handlers(struct ieee80211_rx_data *rx,
-                                         struct ieee80211_rx_data *rx,
+                                  struct sk_buff_head *frames)
-                                         struct sk_buff *skb,
-                                         struct ieee80211_rate *rate)
 {
-        struct sk_buff_head reorder_release;
        ieee80211_rx_result res = RX_DROP_MONITOR;
+        struct sk_buff *skb;
-        __skb_queue_head_init(&reorder_release);
-        rx->skb = skb;
-        rx->sdata = sdata;
 #define CALL_RXH(rxh)                   \
        do {                            \
@@ -2284,23 +2410,14 @@ static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_data *sdata,
                        goto rxh_next;  \
        } while (0);
-        /*
+        while ((skb = __skb_dequeue(frames))) {
-         * NB: the rxh_next label works even if we jump
-         *     to it from here because then the list will
-         *     be empty, which is a trivial check
-         */
-        CALL_RXH(ieee80211_rx_h_passive_scan)
-        CALL_RXH(ieee80211_rx_h_check)
-        ieee80211_rx_reorder_ampdu(rx, &reorder_release);
-        while ((skb = __skb_dequeue(&reorder_release))) {
                /*
                 * all the other fields are valid across frames
                 * that belong to an aMPDU since they are on the
                 * same TID from the same station
                 */
                rx->skb = skb;
+                rx->flags = 0;
                CALL_RXH(ieee80211_rx_h_decrypt)
                CALL_RXH(ieee80211_rx_h_check_more_data)
@@ -2312,50 +2429,92 @@ static void ieee80211_invoke_rx_handlers(struct ieee80211_sub_if_data *sdata,
                CALL_RXH(ieee80211_rx_h_remove_qos_control)
                CALL_RXH(ieee80211_rx_h_amsdu)
 #ifdef CONFIG_MAC80211_MESH
-                if (ieee80211_vif_is_mesh(&sdata->vif))
+                if (ieee80211_vif_is_mesh(&rx->sdata->vif))
                        CALL_RXH(ieee80211_rx_h_mesh_fwding);
 #endif
                CALL_RXH(ieee80211_rx_h_data)
                /* special treatment -- needs the queue */
-                res = ieee80211_rx_h_ctrl(rx, &reorder_release);
+                res = ieee80211_rx_h_ctrl(rx, frames);
                if (res != RX_CONTINUE)
                        goto rxh_next;
+                CALL_RXH(ieee80211_rx_h_mgmt_check)
                CALL_RXH(ieee80211_rx_h_action)
+                CALL_RXH(ieee80211_rx_h_userspace_mgmt)
+                CALL_RXH(ieee80211_rx_h_action_return)
                CALL_RXH(ieee80211_rx_h_mgmt)
+ rxh_next:
+                ieee80211_rx_handlers_result(rx, res);
 #undef CALL_RXH
+        }
+}
+static void ieee80211_invoke_rx_handlers(struct ieee80211_rx_data *rx)
+{
+        struct sk_buff_head reorder_release;
+        ieee80211_rx_result res = RX_DROP_MONITOR;
+        __skb_queue_head_init(&reorder_release);
+#define CALL_RXH(rxh)                   \
+        do {                            \
+                res = rxh(rx);          \
+                if (res != RX_CONTINUE) \
+                        goto rxh_next;  \
+        } while (0);
+        CALL_RXH(ieee80211_rx_h_passive_scan)
+        CALL_RXH(ieee80211_rx_h_check)
+        ieee80211_rx_reorder_ampdu(rx, &reorder_release);
+        ieee80211_rx_handlers(rx, &reorder_release);
+        return;
 rxh_next:
-                switch (res) {
+        ieee80211_rx_handlers_result(rx, res);
-                case RX_DROP_MONITOR:
-                        I802_DEBUG_INC(sdata->local->rx_handlers_drop);
+#undef CALL_RXH
-                        if (rx->sta)
+}
-                                rx->sta->rx_dropped++;
-                        /* fall through */
+/*
-                case RX_CONTINUE:
+ * This function makes calls into the RX path. Therefore the
-                        ieee80211_rx_cooked_monitor(rx, rate);
+ * caller must hold the sta_info->lock and everything has to
-                        break;
+ * be under rcu_read_lock protection as well.
-                case RX_DROP_UNUSABLE:
+ */
-                        I802_DEBUG_INC(sdata->local->rx_handlers_drop);
+void ieee80211_release_reorder_timeout(struct sta_info *sta, int tid)
-                        if (rx->sta)
+{
-                                rx->sta->rx_dropped++;
+        struct sk_buff_head frames;
-                        dev_kfree_skb(rx->skb);
+        struct ieee80211_rx_data rx = {
-                        break;
+                .sta = sta,
-                case RX_QUEUED:
+                .sdata = sta->sdata,
-                        I802_DEBUG_INC(sdata->local->rx_handlers_queued);
+                .local = sta->local,
-                        break;
+                .queue = tid,
-                }
+        };
-        }
+        struct tid_ampdu_rx *tid_agg_rx;
+        tid_agg_rx = rcu_dereference(sta->ampdu_mlme.tid_rx[tid]);
+        if (!tid_agg_rx)
+                return;
+        __skb_queue_head_init(&frames);
+        spin_lock(&tid_agg_rx->reorder_lock);
+        ieee80211_sta_reorder_release(&sta->local->hw, tid_agg_rx, &frames);
+        spin_unlock(&tid_agg_rx->reorder_lock);
+        ieee80211_rx_handlers(&rx, &frames);
 }
 /* main receive path */
-static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
+static int prepare_for_handlers(struct ieee80211_rx_data *rx,
-                                struct ieee80211_rx_data *rx,
                                struct ieee80211_hdr *hdr)
 {
+        struct ieee80211_sub_if_data *sdata = rx->sdata;
        struct sk_buff *skb = rx->skb;
        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
        u8 *bssid = ieee80211_get_bssid(hdr, skb->len, sdata->vif.type);
@@ -2369,7 +2528,7 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
                    compare_ether_addr(sdata->vif.addr, hdr->addr1) != 0) {
                        if (!(sdata->dev->flags & IFF_PROMISC))
                                return 0;
-                        rx->flags &= ~IEEE80211_RX_RA_MATCH;
+                        status->rx_flags &= ~IEEE80211_RX_RA_MATCH;
                }
                break;
        case NL80211_IFTYPE_ADHOC:
@@ -2379,15 +2538,15 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
                        return 1;
                }
                else if (!ieee80211_bssid_match(bssid, sdata->u.ibss.bssid)) {
-                        if (!(rx->flags & IEEE80211_RX_IN_SCAN))
+                        if (!(status->rx_flags & IEEE80211_RX_IN_SCAN))
                                return 0;
-                        rx->flags &= ~IEEE80211_RX_RA_MATCH;
+                        status->rx_flags &= ~IEEE80211_RX_RA_MATCH;
                } else if (!multicast &&
                           compare_ether_addr(sdata->vif.addr,
                                              hdr->addr1) != 0) {
                        if (!(sdata->dev->flags & IFF_PROMISC))
                                return 0;
-                        rx->flags &= ~IEEE80211_RX_RA_MATCH;
+                        status->rx_flags &= ~IEEE80211_RX_RA_MATCH;
                } else if (!rx->sta) {
                        int rate_idx;
                        if (status->flag & RX_FLAG_HT)
@@ -2405,7 +2564,7 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
                        if (!(sdata->dev->flags & IFF_PROMISC))
                                return 0;
-                        rx->flags &= ~IEEE80211_RX_RA_MATCH;
+                        status->rx_flags &= ~IEEE80211_RX_RA_MATCH;
                }
                break;
        case NL80211_IFTYPE_AP_VLAN:
@@ -2416,9 +2575,9 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
                                return 0;
                } else if (!ieee80211_bssid_match(bssid,
                                        sdata->vif.addr)) {
-                        if (!(rx->flags & IEEE80211_RX_IN_SCAN))
+                        if (!(status->rx_flags & IEEE80211_RX_IN_SCAN))
                                return 0;
-                        rx->flags &= ~IEEE80211_RX_RA_MATCH;
+                        status->rx_flags &= ~IEEE80211_RX_RA_MATCH;
                }
                break;
        case NL80211_IFTYPE_WDS:
@@ -2427,9 +2586,7 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
                if (compare_ether_addr(sdata->u.wds.remote_addr, hdr->addr2))
                        return 0;
                break;
-        case NL80211_IFTYPE_MONITOR:
+        default:
-        case NL80211_IFTYPE_UNSPECIFIED:
-        case __NL80211_IFTYPE_AFTER_LAST:
                /* should never get here */
                WARN_ON(1);
                break;
@@ -2439,12 +2596,56 @@ static int prepare_for_handlers(struct ieee80211_sub_if_data *sdata,
 }
 /*
+ * This function returns whether or not the SKB
+ * was destined for RX processing or not, which,
+ * if consume is true, is equivalent to whether
+ * or not the skb was consumed.
+ */
+static bool ieee80211_prepare_and_rx_handle(struct ieee80211_rx_data *rx,
+                                            struct sk_buff *skb, bool consume)
+{
+        struct ieee80211_local *local = rx->local;
+        struct ieee80211_sub_if_data *sdata = rx->sdata;
+        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
+        struct ieee80211_hdr *hdr = (void *)skb->data;
+        int prepares;
+        rx->skb = skb;
+        status->rx_flags |= IEEE80211_RX_RA_MATCH;
+        prepares = prepare_for_handlers(rx, hdr);
+        if (!prepares)
+                return false;
+        if (status->flag & RX_FLAG_MMIC_ERROR) {
+                if (status->rx_flags & IEEE80211_RX_RA_MATCH)
+                        ieee80211_rx_michael_mic_report(hdr, rx);
+                return false;
+        }
+        if (!consume) {
+                skb = skb_copy(skb, GFP_ATOMIC);
+                if (!skb) {
+                        if (net_ratelimit())
+                                wiphy_debug(local->hw.wiphy,
+                                        "failed to copy multicast frame for %s\n",
+                                        sdata->name);
+                        return true;
+                }
+                rx->skb = skb;
+        }
+        ieee80211_invoke_rx_handlers(rx);
+        return true;
+}
+/*
 * This is the actual Rx frames handler. as it blongs to Rx path it must
 * be called with rcu_read_lock protection.
 */
 static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw,
-                                         struct sk_buff *skb,
+                                         struct sk_buff *skb)
-                                         struct ieee80211_rate *rate)
 {
        struct ieee80211_rx_status *status = IEEE80211_SKB_RXCB(skb);
        struct ieee80211_local *local = hw_to_local(hw);
@@ -2452,11 +2653,8 @@ static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw,
        struct ieee80211_hdr *hdr;
        __le16 fc;
        struct ieee80211_rx_data rx;
-        int prepares;
+        struct ieee80211_sub_if_data *prev;
-        struct ieee80211_sub_if_data *prev = NULL;
+        struct sta_info *sta, *tmp, *prev_sta;
-        struct sk_buff *skb_new;
-        struct sta_info *sta, *tmp;
-        bool found_sta = false;
        int err = 0;
        fc = ((struct ieee80211_hdr *)skb->data)->frame_control;
@@ -2469,7 +2667,7 @@ static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw,
        if (unlikely(test_bit(SCAN_HW_SCANNING, &local->scanning) ||
                     test_bit(SCAN_OFF_CHANNEL, &local->scanning)))
-                rx.flags |= IEEE80211_RX_IN_SCAN;
+                status->rx_flags |= IEEE80211_RX_IN_SCAN;
        if (ieee80211_is_mgmt(fc))
                err = skb_linearize(skb);
@@ -2486,91 +2684,67 @@ static void __ieee80211_rx_handle_packet(struct ieee80211_hw *hw,
        ieee80211_verify_alignment(&rx);
        if (ieee80211_is_data(fc)) {
+                prev_sta = NULL;
                for_each_sta_info(local, hdr->addr2, sta, tmp) {
-                        rx.sta = sta;
+                        if (!prev_sta) {
-                        found_sta = true;
+                                prev_sta = sta;
-                        rx.sdata = sta->sdata;
-                        rx.flags |= IEEE80211_RX_RA_MATCH;
-                        prepares = prepare_for_handlers(rx.sdata, &rx, hdr);
-                        if (prepares) {
-                                if (status->flag & RX_FLAG_MMIC_ERROR) {
-                                        if (rx.flags & IEEE80211_RX_RA_MATCH)
-                                                ieee80211_rx_michael_mic_report(hdr, &rx);
-                                } else
-                                        prev = rx.sdata;
-                        }
-                }
-        }
-        if (!found_sta) {
-                list_for_each_entry_rcu(sdata, &local->interfaces, list) {
-                        if (!ieee80211_sdata_running(sdata))
                                continue;
+                        }
-                        if (sdata->vif.type == NL80211_IFTYPE_MONITOR ||
+                        rx.sta = prev_sta;
-                            sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
+                        rx.sdata = prev_sta->sdata;
-                                continue;
+                        ieee80211_prepare_and_rx_handle(&rx, skb, false);
-                        /*
+                        prev_sta = sta;
-                         * frame is destined for this interface, but if it's
+                }
-                         * not also for the previous one we handle that after
-                         * the loop to avoid copying the SKB once too much
-                         */
-                        if (!prev) {
+                if (prev_sta) {
-                                prev = sdata;
+                        rx.sta = prev_sta;
-                                continue;
+                        rx.sdata = prev_sta->sdata;
-                        }
-                        rx.sta = sta_info_get_bss(prev, hdr->addr2);
+                        if (ieee80211_prepare_and_rx_handle(&rx, skb, true))
+                                return;
+                }
+        }
-                        rx.flags |= IEEE80211_RX_RA_MATCH;
+        prev = NULL;
-                        prepares = prepare_for_handlers(prev, &rx, hdr);
-                        if (!prepares)
+        list_for_each_entry_rcu(sdata, &local->interfaces, list) {
-                                goto next;
+                if (!ieee80211_sdata_running(sdata))
+                        continue;
-                        if (status->flag & RX_FLAG_MMIC_ERROR) {
+                if (sdata->vif.type == NL80211_IFTYPE_MONITOR ||
-                                rx.sdata = prev;
+                    sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
-                                if (rx.flags & IEEE80211_RX_RA_MATCH)
+                        continue;
-                                        ieee80211_rx_michael_mic_report(hdr,
-                                                                        &rx);
-                                goto next;
-                        }
-                        /*
+                /*
-                         * frame was destined for the previous interface
+                 * frame is destined for this interface, but if it's
-                         * so invoke RX handlers for it
+                 * not also for the previous one we handle that after
-                         */
+                 * the loop to avoid copying the SKB once too much
+                 */
-                        skb_new = skb_copy(skb, GFP_ATOMIC);
+                if (!prev) {
-                        if (!skb_new) {
-                                if (net_ratelimit())
-                                        printk(KERN_DEBUG "%s: failed to copy "
-                                               "multicast frame for %s\n",
-                                               wiphy_name(local->hw.wiphy),
-                                               prev->name);
-                                goto next;
-                        }
-                        ieee80211_invoke_rx_handlers(prev, &rx, skb_new, rate);
-next:
                        prev = sdata;
+                        continue;
                }
-                if (prev) {
+                rx.sta = sta_info_get_bss(prev, hdr->addr2);
-                        rx.sta = sta_info_get_bss(prev, hdr->addr2);
+                rx.sdata = prev;
+                ieee80211_prepare_and_rx_handle(&rx, skb, false);
-                        rx.flags |= IEEE80211_RX_RA_MATCH;
+                prev = sdata;
-                        prepares = prepare_for_handlers(prev, &rx, hdr);
+        }
-                        if (!prepares)
+        if (prev) {
-                                prev = NULL;
+                rx.sta = sta_info_get_bss(prev, hdr->addr2);
-                }
+                rx.sdata = prev;
+                if (ieee80211_prepare_and_rx_handle(&rx, skb, true))
+                        return;
        }
-        if (prev)
-                ieee80211_invoke_rx_handlers(prev, &rx, skb, rate);
+        dev_kfree_skb(skb);
-        else
-                dev_kfree_skb(skb);
 }
 /*
@@ -2611,30 +2785,41 @@ void ieee80211_rx(struct ieee80211_hw *hw, struct sk_buff *skb)
        if (WARN_ON(!local->started))
                goto drop;
-        if (status->flag & RX_FLAG_HT) {
+        if (likely(!(status->flag & RX_FLAG_FAILED_PLCP_CRC))) {
                /*
-                 * rate_idx is MCS index, which can be [0-76] as documented on:
+                 * Validate the rate, unless a PLCP error means that
-                 *
+                 * we probably can't have a valid rate here anyway.
-                 * http://wireless.kernel.org/en/developers/Documentation/ieee80211/802.11n
-                 *
-                 * Anything else would be some sort of driver or hardware error.
-                 * The driver should catch hardware errors.
                 */
-                if (WARN((status->rate_idx < 0 ||
-                         status->rate_idx > 76),
+                if (status->flag & RX_FLAG_HT) {
-                         "Rate marked as an HT rate but passed "
+                        /*
-                         "status->rate_idx is not "
+                         * rate_idx is MCS index, which can be [0-76]
-                         "an MCS index [0-76]: %d (0x%02x)\n",
+                         * as documented on:
-                         status->rate_idx,
+                         *
-                         status->rate_idx))
+                         * http://wireless.kernel.org/en/developers/Documentation/ieee80211/802.11n
-                        goto drop;
+                         *
-        } else {
+                         * Anything else would be some sort of driver or
-                if (WARN_ON(status->rate_idx < 0 ||
+                         * hardware error. The driver should catch hardware
-                            status->rate_idx >= sband->n_bitrates))
+                         * errors.
-                        goto drop;
+                         */
-                rate = &sband->bitrates[status->rate_idx];
+                        if (WARN((status->rate_idx < 0 ||
+                                 status->rate_idx > 76),
+                                 "Rate marked as an HT rate but passed "
+                                 "status->rate_idx is not "
+                                 "an MCS index [0-76]: %d (0x%02x)\n",
+                                 status->rate_idx,
+                                 status->rate_idx))
+                                goto drop;
+                } else {
+                        if (WARN_ON(status->rate_idx < 0 ||
+                                    status->rate_idx >= sband->n_bitrates))
+                                goto drop;
+                        rate = &sband->bitrates[status->rate_idx];
+                }
        }
+        status->rx_flags = 0;
        /*
         * key references and virtual interfaces are protected using RCU
         * and this requires that we are in a read-side RCU section during
@@ -2654,7 +2839,7 @@ void ieee80211_rx(struct ieee80211_hw *hw, struct sk_buff *skb)
                return;
        }
-        __ieee80211_rx_handle_packet(hw, skb, rate);
+        __ieee80211_rx_handle_packet(hw, skb);
        rcu_read_unlock();
diff --git a/net/mac80211/scan.c b/net/mac80211/scan.c
index 872d7b6ef6b3..5171a9581631 100644
--- a/net/mac80211/scan.c
+++ b/net/mac80211/scan.c
@@ -242,20 +242,19 @@ static bool ieee80211_prep_hw_scan(struct ieee80211_local *local)
        local->hw_scan_req->n_channels = n_chans;
        ielen = ieee80211_build_preq_ies(local, (u8 *)local->hw_scan_req->ie,
-                                         req->ie, req->ie_len, band);
+                                         req->ie, req->ie_len, band, (u32) -1,
+                                         0);
        local->hw_scan_req->ie_len = ielen;
        return true;
 }
-void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
+static void __ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
 {
        struct ieee80211_local *local = hw_to_local(hw);
        bool was_hw_scan;
-        trace_api_scan_completed(local, aborted);
+        mutex_lock(&local->mtx);
-        mutex_lock(&local->scan_mtx);
        /*
         * It's ok to abort a not-yet-running scan (that
@@ -267,7 +266,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
                aborted = true;
        if (WARN_ON(!local->scan_req)) {
-                mutex_unlock(&local->scan_mtx);
+                mutex_unlock(&local->mtx);
                return;
        }
@@ -275,7 +274,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
        if (was_hw_scan && !aborted && ieee80211_prep_hw_scan(local)) {
                ieee80211_queue_delayed_work(&local->hw,
                                             &local->scan_work, 0);
-                mutex_unlock(&local->scan_mtx);
+                mutex_unlock(&local->mtx);
                return;
        }
@@ -291,7 +290,7 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
        local->scan_channel = NULL;
        /* we only have to protect scan_req and hw/sw scan */
-        mutex_unlock(&local->scan_mtx);
+        mutex_unlock(&local->mtx);
        ieee80211_hw_config(local, IEEE80211_CONF_CHANGE_CHANNEL);
        if (was_hw_scan)
@@ -304,12 +303,26 @@ void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
        ieee80211_offchannel_return(local, true);
 done:
+        mutex_lock(&local->mtx);
        ieee80211_recalc_idle(local);
+        mutex_unlock(&local->mtx);
        ieee80211_mlme_notify_scan_completed(local);
        ieee80211_ibss_notify_scan_completed(local);
        ieee80211_mesh_notify_scan_completed(local);
        ieee80211_queue_work(&local->hw, &local->work_work);
 }
+void ieee80211_scan_completed(struct ieee80211_hw *hw, bool aborted)
+{
+        struct ieee80211_local *local = hw_to_local(hw);
+        trace_api_scan_completed(local, aborted);
+        set_bit(SCAN_COMPLETED, &local->scanning);
+        if (aborted)
+                set_bit(SCAN_ABORTED, &local->scanning);
+        ieee80211_queue_delayed_work(&local->hw, &local->scan_work, 0);
+}
 EXPORT_SYMBOL(ieee80211_scan_completed);
 static int ieee80211_start_sw_scan(struct ieee80211_local *local)
@@ -447,7 +460,7 @@ static int ieee80211_scan_state_decision(struct ieee80211_local *local,
        /* if no more bands/channels left, complete scan and advance to the idle state */
        if (local->scan_channel_idx >= local->scan_req->n_channels) {
-                ieee80211_scan_completed(&local->hw, false);
+                __ieee80211_scan_completed(&local->hw, false);
                return 1;
        }
@@ -639,17 +652,25 @@ void ieee80211_scan_work(struct work_struct *work)
        struct ieee80211_sub_if_data *sdata = local->scan_sdata;
        unsigned long next_delay = 0;
-        mutex_lock(&local->scan_mtx);
+        if (test_and_clear_bit(SCAN_COMPLETED, &local->scanning)) {
+                bool aborted;
+                aborted = test_and_clear_bit(SCAN_ABORTED, &local->scanning);
+                __ieee80211_scan_completed(&local->hw, aborted);
+                return;
+        }
+        mutex_lock(&local->mtx);
        if (!sdata || !local->scan_req) {
-                mutex_unlock(&local->scan_mtx);
+                mutex_unlock(&local->mtx);
                return;
        }
        if (local->hw_scan_req) {
                int rc = drv_hw_scan(local, sdata, local->hw_scan_req);
-                mutex_unlock(&local->scan_mtx);
+                mutex_unlock(&local->mtx);
                if (rc)
-                        ieee80211_scan_completed(&local->hw, true);
+                        __ieee80211_scan_completed(&local->hw, true);
                return;
        }
@@ -661,20 +682,20 @@ void ieee80211_scan_work(struct work_struct *work)
                local->scan_sdata = NULL;
                rc = __ieee80211_start_scan(sdata, req);
-                mutex_unlock(&local->scan_mtx);
+                mutex_unlock(&local->mtx);
                if (rc)
-                        ieee80211_scan_completed(&local->hw, true);
+                        __ieee80211_scan_completed(&local->hw, true);
                return;
        }
-        mutex_unlock(&local->scan_mtx);
+        mutex_unlock(&local->mtx);
        /*
         * Avoid re-scheduling when the sdata is going away.
         */
        if (!ieee80211_sdata_running(sdata)) {
-                ieee80211_scan_completed(&local->hw, true);
+                __ieee80211_scan_completed(&local->hw, true);
                return;
        }
@@ -711,9 +732,9 @@ int ieee80211_request_scan(struct ieee80211_sub_if_data *sdata,
 {
        int res;
-        mutex_lock(&sdata->local->scan_mtx);
+        mutex_lock(&sdata->local->mtx);
        res = __ieee80211_start_scan(sdata, req);
-        mutex_unlock(&sdata->local->scan_mtx);
+        mutex_unlock(&sdata->local->mtx);
        return res;
 }
@@ -726,7 +747,7 @@ int ieee80211_request_internal_scan(struct ieee80211_sub_if_data *sdata,
        int ret = -EBUSY;
        enum ieee80211_band band;
-        mutex_lock(&local->scan_mtx);
+        mutex_lock(&local->mtx);
        /* busy scanning */
        if (local->scan_req)
@@ -761,7 +782,7 @@ int ieee80211_request_internal_scan(struct ieee80211_sub_if_data *sdata,
        ret = __ieee80211_start_scan(sdata, sdata->local->int_scan_req);
 unlock:
-        mutex_unlock(&local->scan_mtx);
+        mutex_unlock(&local->mtx);
        return ret;
 }
@@ -775,11 +796,11 @@ void ieee80211_scan_cancel(struct ieee80211_local *local)
         * Only call this function when a scan can't be
         * queued -- mostly at suspend under RTNL.
         */
-        mutex_lock(&local->scan_mtx);
+        mutex_lock(&local->mtx);
        abortscan = test_bit(SCAN_SW_SCANNING, &local->scanning) ||
                    (!local->scanning && local->scan_req);
-        mutex_unlock(&local->scan_mtx);
+        mutex_unlock(&local->mtx);
        if (abortscan)
-                ieee80211_scan_completed(&local->hw, true);
+                __ieee80211_scan_completed(&local->hw, true);
 }
diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c
index 6d86f0c1ad04..ca2cba9cea87 100644
--- a/net/mac80211/sta_info.c
+++ b/net/mac80211/sta_info.c
@@ -125,7 +125,7 @@ struct sta_info *sta_info_get_bss(struct ieee80211_sub_if_data *sdata,
                                    lockdep_is_held(&local->sta_mtx));
        while (sta) {
                if ((sta->sdata == sdata ||
-                     sta->sdata->bss == sdata->bss) &&
+                     (sta->sdata->bss && sta->sdata->bss == sdata->bss)) &&
                    memcmp(sta->sta.addr, addr, ETH_ALEN) == 0)
                        break;
                sta = rcu_dereference_check(sta->hnext,
@@ -174,8 +174,7 @@ static void __sta_info_free(struct ieee80211_local *local,
        }
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: Destroyed STA %pM\n",
+        wiphy_debug(local->hw.wiphy, "Destroyed STA %pM\n", sta->sta.addr);
-               wiphy_name(local->hw.wiphy), sta->sta.addr);
 #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */
        kfree(sta);
@@ -262,8 +261,7 @@ struct sta_info *sta_info_alloc(struct ieee80211_sub_if_data *sdata,
                sta->last_seq_ctrl[i] = cpu_to_le16(USHRT_MAX);
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: Allocated STA %pM\n",
+        wiphy_debug(local->hw.wiphy, "Allocated STA %pM\n", sta->sta.addr);
-               wiphy_name(local->hw.wiphy), sta->sta.addr);
 #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */
 #ifdef CONFIG_MAC80211_MESH
@@ -282,7 +280,7 @@ static int sta_info_finish_insert(struct sta_info *sta, bool async)
        unsigned long flags;
        int err = 0;
-        WARN_ON(!mutex_is_locked(&local->sta_mtx));
+        lockdep_assert_held(&local->sta_mtx);
        /* notify driver */
        if (sdata->vif.type == NL80211_IFTYPE_AP_VLAN)
@@ -300,8 +298,9 @@ static int sta_info_finish_insert(struct sta_info *sta, bool async)
                sta->uploaded = true;
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
                if (async)
-                        printk(KERN_DEBUG "%s: Finished adding IBSS STA %pM\n",
+                        wiphy_debug(local->hw.wiphy,
-                               wiphy_name(local->hw.wiphy), sta->sta.addr);
+                                    "Finished adding IBSS STA %pM\n",
+                                    sta->sta.addr);
 #endif
        }
@@ -411,8 +410,8 @@ int sta_info_insert_rcu(struct sta_info *sta) __acquires(RCU)
                spin_unlock_irqrestore(&local->sta_lock, flags);
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-                printk(KERN_DEBUG "%s: Added IBSS STA %pM\n",
+                wiphy_debug(local->hw.wiphy, "Added IBSS STA %pM\n",
-                       wiphy_name(local->hw.wiphy), sta->sta.addr);
+                            sta->sta.addr);
 #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */
                ieee80211_queue_work(&local->hw, &local->sta_finish_work);
@@ -459,8 +458,7 @@ int sta_info_insert_rcu(struct sta_info *sta) __acquires(RCU)
        }
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: Inserted STA %pM\n",
+        wiphy_debug(local->hw.wiphy, "Inserted STA %pM\n", sta->sta.addr);
-               wiphy_name(local->hw.wiphy), sta->sta.addr);
 #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */
        /* move reference to rcu-protected */
@@ -690,8 +688,7 @@ static int __must_check __sta_info_destroy(struct sta_info *sta)
 #endif
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
-        printk(KERN_DEBUG "%s: Removed STA %pM\n",
+        wiphy_debug(local->hw.wiphy, "Removed STA %pM\n", sta->sta.addr);
-               wiphy_name(local->hw.wiphy), sta->sta.addr);
 #endif /* CONFIG_MAC80211_VERBOSE_DEBUG */
        cancel_work_sync(&sta->drv_unblock_wk);
@@ -841,13 +838,20 @@ void ieee80211_sta_expire(struct ieee80211_sub_if_data *sdata,
        mutex_unlock(&local->sta_mtx);
 }
-struct ieee80211_sta *ieee80211_find_sta_by_hw(struct ieee80211_hw *hw,
+struct ieee80211_sta *ieee80211_find_sta_by_ifaddr(struct ieee80211_hw *hw,
-                                               const u8 *addr)
+                                               const u8 *addr,
+                                               const u8 *localaddr)
 {
        struct sta_info *sta, *nxt;
-        /* Just return a random station ... first in list ... */
+        /*
+         * Just return a random station if localaddr is NULL
+         * ... first in list.
+         */
        for_each_sta_info(hw_to_local(hw), addr, sta, nxt) {
+                if (localaddr &&
+                    compare_ether_addr(sta->sdata->vif.addr, localaddr) != 0)
+                        continue;
                if (!sta->uploaded)
                        return NULL;
                return &sta->sta;
@@ -855,7 +859,7 @@ struct ieee80211_sta *ieee80211_find_sta_by_hw(struct ieee80211_hw *hw,
        return NULL;
 }
-EXPORT_SYMBOL_GPL(ieee80211_find_sta_by_hw);
+EXPORT_SYMBOL_GPL(ieee80211_find_sta_by_ifaddr);
 struct ieee80211_sta *ieee80211_find_sta(struct ieee80211_vif *vif,
                                         const u8 *addr)
diff --git a/net/mac80211/sta_info.h b/net/mac80211/sta_info.h
index 54262e72376d..810c5ce98316 100644
--- a/net/mac80211/sta_info.h
+++ b/net/mac80211/sta_info.h
@@ -103,6 +103,7 @@ struct tid_ampdu_tx {
 * @reorder_buf: buffer to reorder incoming aggregated MPDUs
 * @reorder_time: jiffies when skb was added
 * @session_timer: check if peer keeps Tx-ing on the TID (by timeout value)
+ * @reorder_timer: releases expired frames from the reorder buffer.
 * @head_seq_num: head sequence number in reordering buffer.
 * @stored_mpdu_num: number of MPDUs in reordering buffer
 * @ssn: Starting Sequence Number expected to be aggregated.
@@ -110,20 +111,25 @@ struct tid_ampdu_tx {
 * @timeout: reset timer value (in TUs).
 * @dialog_token: dialog token for aggregation session
 * @rcu_head: RCU head used for freeing this struct
+ * @reorder_lock: serializes access to reorder buffer, see below.
 *
 * This structure is protected by RCU and the per-station
 * spinlock. Assignments to the array holding it must hold
- * the spinlock, only the RX path can access it under RCU
+ * the spinlock.
- * lock-free. The RX path, since it is single-threaded,
+ *
- * can even modify the structure without locking since the
+ * The @reorder_lock is used to protect the variables and
- * only other modifications to it are done when the struct
+ * arrays such as @reorder_buf, @reorder_time, @head_seq_num,
- * can not yet or no longer be found by the RX path.
+ * @stored_mpdu_num and @reorder_time from being corrupted by
+ * concurrent access of the RX path and the expired frame
+ * release timer.
 */
 struct tid_ampdu_rx {
        struct rcu_head rcu_head;
+        spinlock_t reorder_lock;
        struct sk_buff **reorder_buf;
        unsigned long *reorder_time;
        struct timer_list session_timer;
+        struct timer_list reorder_timer;
        u16 head_seq_num;
        u16 stored_mpdu_num;
        u16 ssn;
diff --git a/net/mac80211/status.c b/net/mac80211/status.c
index 10caec5ea8fa..dd85006c4fe8 100644
--- a/net/mac80211/status.c
+++ b/net/mac80211/status.c
@@ -58,6 +58,7 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local,
        info->control.vif = &sta->sdata->vif;
        info->flags |= IEEE80211_TX_INTFL_NEED_TXPROCESSING |
                       IEEE80211_TX_INTFL_RETRANSMISSION;
+        info->flags &= ~IEEE80211_TX_TEMPORARY_FLAGS;
        sta->tx_filtered_count++;
@@ -114,11 +115,10 @@ static void ieee80211_handle_filtered_frame(struct ieee80211_local *local,
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
        if (net_ratelimit())
-                printk(KERN_DEBUG "%s: dropped TX filtered frame, "
+                wiphy_debug(local->hw.wiphy,
-                       "queue_len=%d PS=%d @%lu\n",
+                            "dropped TX filtered frame, queue_len=%d PS=%d @%lu\n",
-                       wiphy_name(local->hw.wiphy),
+                            skb_queue_len(&sta->tx_filtered),
-                       skb_queue_len(&sta->tx_filtered),
+                            !!test_sta_flags(sta, WLAN_STA_PS_STA), jiffies);
-                       !!test_sta_flags(sta, WLAN_STA_PS_STA), jiffies);
 #endif
        dev_kfree_skb(skb);
 }
@@ -296,7 +296,7 @@ void ieee80211_tx_status(struct ieee80211_hw *hw, struct sk_buff *skb)
        }
        if (info->flags & IEEE80211_TX_INTFL_NL80211_FRAME_TX)
-                cfg80211_action_tx_status(
+                cfg80211_mgmt_tx_status(
                        skb->dev, (unsigned long) skb, skb->data, skb->len,
                        !!(info->flags & IEEE80211_TX_STAT_ACK), GFP_ATOMIC);
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index c54db966926b..e1733dcb58a7 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -351,8 +351,8 @@ static void purge_old_ps_buffers(struct ieee80211_local *local)
        local->total_ps_buffered = total;
 #ifdef CONFIG_MAC80211_VERBOSE_PS_DEBUG
-        printk(KERN_DEBUG "%s: PS buffers full - purged %d frames\n",
+        wiphy_debug(local->hw.wiphy, "PS buffers full - purged %d frames\n",
-               wiphy_name(local->hw.wiphy), purged);
+                    purged);
 #endif
 }
@@ -509,6 +509,18 @@ ieee80211_tx_h_ps_buf(struct ieee80211_tx_data *tx)
 }
 static ieee80211_tx_result debug_noinline
+ieee80211_tx_h_check_control_port_protocol(struct ieee80211_tx_data *tx)
+{
+        struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb);
+        if (unlikely(tx->sdata->control_port_protocol == tx->skb->protocol &&
+                     tx->sdata->control_port_no_encrypt))
+                info->flags |= IEEE80211_TX_INTFL_DONT_ENCRYPT;
+        return TX_CONTINUE;
+}
+static ieee80211_tx_result debug_noinline
 ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx)
 {
        struct ieee80211_key *key = NULL;
@@ -527,7 +539,7 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx)
        else if ((key = rcu_dereference(tx->sdata->default_key)))
                tx->key = key;
        else if (tx->sdata->drop_unencrypted &&
-                 (tx->skb->protocol != cpu_to_be16(ETH_P_PAE)) &&
+                 (tx->skb->protocol != tx->sdata->control_port_protocol) &&
                 !(info->flags & IEEE80211_TX_CTL_INJECTED) &&
                 (!ieee80211_is_robust_mgmt_frame(hdr) ||
                  (ieee80211_is_action(hdr->frame_control) &&
@@ -543,15 +555,16 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx)
                tx->key->tx_rx_count++;
                /* TODO: add threshold stuff again */
-                switch (tx->key->conf.alg) {
+                switch (tx->key->conf.cipher) {
-                case ALG_WEP:
+                case WLAN_CIPHER_SUITE_WEP40:
+                case WLAN_CIPHER_SUITE_WEP104:
                        if (ieee80211_is_auth(hdr->frame_control))
                                break;
-                case ALG_TKIP:
+                case WLAN_CIPHER_SUITE_TKIP:
                        if (!ieee80211_is_data_present(hdr->frame_control))
                                tx->key = NULL;
                        break;
-                case ALG_CCMP:
+                case WLAN_CIPHER_SUITE_CCMP:
                        if (!ieee80211_is_data_present(hdr->frame_control) &&
                            !ieee80211_use_mfp(hdr->frame_control, tx->sta,
                                               tx->skb))
@@ -561,7 +574,7 @@ ieee80211_tx_h_select_key(struct ieee80211_tx_data *tx)
                                           IEEE80211_KEY_FLAG_SW_MGMT) &&
                                        ieee80211_is_mgmt(hdr->frame_control);
                        break;
-                case ALG_AES_CMAC:
+                case WLAN_CIPHER_SUITE_AES_CMAC:
                        if (!ieee80211_is_mgmt(hdr->frame_control))
                                tx->key = NULL;
                        break;
@@ -946,22 +959,31 @@ ieee80211_tx_h_stats(struct ieee80211_tx_data *tx)
 static ieee80211_tx_result debug_noinline
 ieee80211_tx_h_encrypt(struct ieee80211_tx_data *tx)
 {
+        struct ieee80211_tx_info *info = IEEE80211_SKB_CB(tx->skb);
        if (!tx->key)
                return TX_CONTINUE;
-        switch (tx->key->conf.alg) {
+        switch (tx->key->conf.cipher) {
-        case ALG_WEP:
+        case WLAN_CIPHER_SUITE_WEP40:
+        case WLAN_CIPHER_SUITE_WEP104:
                return ieee80211_crypto_wep_encrypt(tx);
-        case ALG_TKIP:
+        case WLAN_CIPHER_SUITE_TKIP:
                return ieee80211_crypto_tkip_encrypt(tx);
-        case ALG_CCMP:
+        case WLAN_CIPHER_SUITE_CCMP:
                return ieee80211_crypto_ccmp_encrypt(tx);
-        case ALG_AES_CMAC:
+        case WLAN_CIPHER_SUITE_AES_CMAC:
                return ieee80211_crypto_aes_cmac_encrypt(tx);
+        default:
+                /* handle hw-only algorithm */
+                if (info->control.hw_key) {
+                        ieee80211_tx_set_protected(tx);
+                        return TX_CONTINUE;
+                }
+                break;
        }
-        /* not reached */
-        WARN_ON(1);
        return TX_DROP;
 }
@@ -1339,6 +1361,7 @@ static int invoke_tx_handlers(struct ieee80211_tx_data *tx)
        CALL_TXH(ieee80211_tx_h_dynamic_ps);
        CALL_TXH(ieee80211_tx_h_check_assoc);
        CALL_TXH(ieee80211_tx_h_ps_buf);
+        CALL_TXH(ieee80211_tx_h_check_control_port_protocol);
        CALL_TXH(ieee80211_tx_h_select_key);
        if (!(tx->local->hw.flags & IEEE80211_HW_HAS_RATE_CONTROL))
                CALL_TXH(ieee80211_tx_h_rate_ctrl);
@@ -1511,8 +1534,8 @@ static int ieee80211_skb_resize(struct ieee80211_local *local,
                I802_DEBUG_INC(local->tx_expand_skb_head);
        if (pskb_expand_head(skb, head_need, tail_need, GFP_ATOMIC)) {
-                printk(KERN_DEBUG "%s: failed to reallocate TX buffer\n",
+                wiphy_debug(local->hw.wiphy,
-                       wiphy_name(local->hw.wiphy));
+                            "failed to reallocate TX buffer\n");
                return -ENOMEM;
        }
@@ -1586,6 +1609,7 @@ static void ieee80211_xmit(struct ieee80211_sub_if_data *sdata,
                return;
        }
+        hdr = (struct ieee80211_hdr *) skb->data;
        info->control.vif = &sdata->vif;
        if (ieee80211_vif_is_mesh(&sdata->vif) &&
@@ -1699,7 +1723,7 @@ netdev_tx_t ieee80211_subif_start_xmit(struct sk_buff *skb,
        u16 ethertype, hdrlen,  meshhdrlen = 0;
        __le16 fc;
        struct ieee80211_hdr hdr;
-        struct ieee80211s_hdr mesh_hdr;
+        struct ieee80211s_hdr mesh_hdr __maybe_unused;
        const u8 *encaps_data;
        int encaps_len, skip_header_bytes;
        int nh_pos, h_pos;
@@ -1816,7 +1840,8 @@ netdev_tx_t ieee80211_subif_start_xmit(struct sk_buff *skb,
 #endif
        case NL80211_IFTYPE_STATION:
                memcpy(hdr.addr1, sdata->u.mgd.bssid, ETH_ALEN);
-                if (sdata->u.mgd.use_4addr && ethertype != ETH_P_PAE) {
+                if (sdata->u.mgd.use_4addr &&
+                    cpu_to_be16(ethertype) != sdata->control_port_protocol) {
                        fc |= cpu_to_le16(IEEE80211_FCTL_FROMDS | IEEE80211_FCTL_TODS);
                        /* RA TA DA SA */
                        memcpy(hdr.addr2, sdata->vif.addr, ETH_ALEN);
@@ -1869,7 +1894,7 @@ netdev_tx_t ieee80211_subif_start_xmit(struct sk_buff *skb,
        if (!ieee80211_vif_is_mesh(&sdata->vif) &&
                unlikely(!is_multicast_ether_addr(hdr.addr1) &&
                      !(sta_flags & WLAN_STA_AUTHORIZED) &&
-                      !(ethertype == ETH_P_PAE &&
+                      !(cpu_to_be16(ethertype) == sdata->control_port_protocol &&
                       compare_ether_addr(sdata->vif.addr,
                                          skb->data + ETH_ALEN) == 0))) {
 #ifdef CONFIG_MAC80211_VERBOSE_DEBUG
@@ -2068,8 +2093,7 @@ void ieee80211_tx_pending(unsigned long data)
                if (skb_queue_empty(&local->pending[i]))
                        list_for_each_entry_rcu(sdata, &local->interfaces, list)
-                                netif_tx_wake_queue(
+                                netif_wake_subqueue(sdata->dev, i);
-                                        netdev_get_tx_queue(sdata->dev, i));
        }
        spin_unlock_irqrestore(&local->queue_stop_reason_lock, flags);
diff --git a/net/mac80211/util.c b/net/mac80211/util.c
index 748387d45bc0..aba025d748e9 100644
--- a/net/mac80211/util.c
+++ b/net/mac80211/util.c
@@ -283,8 +283,11 @@ static void __ieee80211_wake_queue(struct ieee80211_hw *hw, int queue,
        if (skb_queue_empty(&local->pending[queue])) {
                rcu_read_lock();
-                list_for_each_entry_rcu(sdata, &local->interfaces, list)
+                list_for_each_entry_rcu(sdata, &local->interfaces, list) {
-                        netif_tx_wake_queue(netdev_get_tx_queue(sdata->dev, queue));
+                        if (test_bit(SDATA_STATE_OFFCHANNEL, &sdata->state))
+                                continue;
+                        netif_wake_subqueue(sdata->dev, queue);
+                }
                rcu_read_unlock();
        } else
                tasklet_schedule(&local->tx_pending_tasklet);
@@ -323,7 +326,7 @@ static void __ieee80211_stop_queue(struct ieee80211_hw *hw, int queue,
        rcu_read_lock();
        list_for_each_entry_rcu(sdata, &local->interfaces, list)
-                netif_tx_stop_queue(netdev_get_tx_queue(sdata->dev, queue));
+                netif_stop_subqueue(sdata->dev, queue);
        rcu_read_unlock();
 }
@@ -471,16 +474,10 @@ void ieee80211_iterate_active_interfaces(
        list_for_each_entry(sdata, &local->interfaces, list) {
                switch (sdata->vif.type) {
-                case __NL80211_IFTYPE_AFTER_LAST:
-                case NL80211_IFTYPE_UNSPECIFIED:
                case NL80211_IFTYPE_MONITOR:
                case NL80211_IFTYPE_AP_VLAN:
                        continue;
-                case NL80211_IFTYPE_AP:
+                default:
-                case NL80211_IFTYPE_STATION:
-                case NL80211_IFTYPE_ADHOC:
-                case NL80211_IFTYPE_WDS:
-                case NL80211_IFTYPE_MESH_POINT:
                        break;
                }
                if (ieee80211_sdata_running(sdata))
@@ -505,16 +502,10 @@ void ieee80211_iterate_active_interfaces_atomic(
        list_for_each_entry_rcu(sdata, &local->interfaces, list) {
                switch (sdata->vif.type) {
-                case __NL80211_IFTYPE_AFTER_LAST:
-                case NL80211_IFTYPE_UNSPECIFIED:
                case NL80211_IFTYPE_MONITOR:
                case NL80211_IFTYPE_AP_VLAN:
                        continue;
-                case NL80211_IFTYPE_AP:
+                default:
-                case NL80211_IFTYPE_STATION:
-                case NL80211_IFTYPE_ADHOC:
-                case NL80211_IFTYPE_WDS:
-                case NL80211_IFTYPE_MESH_POINT:
                        break;
                }
                if (ieee80211_sdata_running(sdata))
@@ -904,26 +895,34 @@ void ieee80211_send_auth(struct ieee80211_sub_if_data *sdata,
 int ieee80211_build_preq_ies(struct ieee80211_local *local, u8 *buffer,
                             const u8 *ie, size_t ie_len,
-                             enum ieee80211_band band)
+                             enum ieee80211_band band, u32 rate_mask,
+                             u8 channel)
 {
        struct ieee80211_supported_band *sband;
        u8 *pos;
        size_t offset = 0, noffset;
        int supp_rates_len, i;
+        u8 rates[32];
+        int num_rates;
+        int ext_rates_len;
        sband = local->hw.wiphy->bands[band];
        pos = buffer;
-        supp_rates_len = min_t(int, sband->n_bitrates, 8);
+        num_rates = 0;
+        for (i = 0; i < sband->n_bitrates; i++) {
+                if ((BIT(i) & rate_mask) == 0)
+                        continue; /* skip rate */
+                rates[num_rates++] = (u8) (sband->bitrates[i].bitrate / 5);
+        }
+        supp_rates_len = min_t(int, num_rates, 8);
        *pos++ = WLAN_EID_SUPP_RATES;
        *pos++ = supp_rates_len;
+        memcpy(pos, rates, supp_rates_len);
-        for (i = 0; i < supp_rates_len; i++) {
+        pos += supp_rates_len;
-                int rate = sband->bitrates[i].bitrate;
-                *pos++ = (u8) (rate / 5);
-        }
        /* insert "request information" if in custom IEs */
        if (ie && ie_len) {
@@ -941,14 +940,18 @@ int ieee80211_build_preq_ies(struct ieee80211_local *local, u8 *buffer,
                offset = noffset;
        }
-        if (sband->n_bitrates > i) {
+        ext_rates_len = num_rates - supp_rates_len;
+        if (ext_rates_len > 0) {
                *pos++ = WLAN_EID_EXT_SUPP_RATES;
-                *pos++ = sband->n_bitrates - i;
+                *pos++ = ext_rates_len;
+                memcpy(pos, rates + supp_rates_len, ext_rates_len);
+                pos += ext_rates_len;
+        }
-                for (; i < sband->n_bitrates; i++) {
+        if (channel && sband->band == IEEE80211_BAND_2GHZ) {
-                        int rate = sband->bitrates[i].bitrate;
+                *pos++ = WLAN_EID_DS_PARAMS;
-                        *pos++ = (u8) (rate / 5);
+                *pos++ = 1;
-                }
+                *pos++ = channel;
        }
        /* insert custom IEs that go before HT */
@@ -1017,6 +1020,7 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
        struct ieee80211_mgmt *mgmt;
        size_t buf_len;
        u8 *buf;
+        u8 chan;
        /* FIXME: come up with a proper value */
        buf = kmalloc(200 + ie_len, GFP_KERNEL);
@@ -1026,8 +1030,14 @@ void ieee80211_send_probe_req(struct ieee80211_sub_if_data *sdata, u8 *dst,
                return;
        }
+        chan = ieee80211_frequency_to_channel(
+                local->hw.conf.channel->center_freq);
        buf_len = ieee80211_build_preq_ies(local, buf, ie, ie_len,
-                                           local->hw.conf.channel->band);
+                                           local->hw.conf.channel->band,
+                                           sdata->rc_rateidx_mask
+                                           [local->hw.conf.channel->band],
+                                           chan);
        skb = ieee80211_probereq_get(&local->hw, &sdata->vif,
                                     ssid, ssid_len,
@@ -1189,7 +1199,9 @@ int ieee80211_reconfig(struct ieee80211_local *local)
                        /* ignore virtual */
                        break;
                case NL80211_IFTYPE_UNSPECIFIED:
-                case __NL80211_IFTYPE_AFTER_LAST:
+                case NUM_NL80211_IFTYPES:
+                case NL80211_IFTYPE_P2P_CLIENT:
+                case NL80211_IFTYPE_P2P_GO:
                        WARN_ON(1);
                        break;
                }
@@ -1293,9 +1305,9 @@ void ieee80211_recalc_smps(struct ieee80211_local *local,
        int count = 0;
        if (forsdata)
-                WARN_ON(!mutex_is_locked(&forsdata->u.mgd.mtx));
+                lockdep_assert_held(&forsdata->u.mgd.mtx);
-        WARN_ON(!mutex_is_locked(&local->iflist_mtx));
+        lockdep_assert_held(&local->iflist_mtx);
        /*
         * This function could be improved to handle multiple
@@ -1308,7 +1320,7 @@ void ieee80211_recalc_smps(struct ieee80211_local *local,
         */
        list_for_each_entry(sdata, &local->interfaces, list) {
-                if (!netif_running(sdata->dev))
+                if (!ieee80211_sdata_running(sdata))
                        continue;
                if (sdata->vif.type != NL80211_IFTYPE_STATION)
                        goto set;
diff --git a/net/mac80211/wep.c b/net/mac80211/wep.c
index 9ebc8d8a1f5b..f27484c22b9f 100644
--- a/net/mac80211/wep.c
+++ b/net/mac80211/wep.c
@@ -240,7 +240,7 @@ static int ieee80211_wep_decrypt(struct ieee80211_local *local,
        keyidx = skb->data[hdrlen + 3] >> 6;
-        if (!key || keyidx != key->conf.keyidx || key->conf.alg != ALG_WEP)
+        if (!key || keyidx != key->conf.keyidx)
                return -1;
        klen = 3 + key->conf.keylen;
diff --git a/net/mac80211/work.c b/net/mac80211/work.c
index 81d4ad64184a..ae344d1ba056 100644
--- a/net/mac80211/work.c
+++ b/net/mac80211/work.c
@@ -43,7 +43,7 @@ enum work_action {
 /* utils */
 static inline void ASSERT_WORK_MTX(struct ieee80211_local *local)
 {
-        WARN_ON(!mutex_is_locked(&local->work_mtx));
+        lockdep_assert_held(&local->mtx);
 }
 /*
@@ -757,7 +757,7 @@ static void ieee80211_work_rx_queued_mgmt(struct ieee80211_local *local,
        mgmt = (struct ieee80211_mgmt *) skb->data;
        fc = le16_to_cpu(mgmt->frame_control);
-        mutex_lock(&local->work_mtx);
+        mutex_lock(&local->mtx);
        list_for_each_entry(wk, &local->work_list, list) {
                const u8 *bssid = NULL;
@@ -833,7 +833,7 @@ static void ieee80211_work_rx_queued_mgmt(struct ieee80211_local *local,
                WARN(1, "unexpected: %d", rma);
        }
-        mutex_unlock(&local->work_mtx);
+        mutex_unlock(&local->mtx);
        if (rma != WORK_ACT_DONE)
                goto out;
@@ -845,9 +845,9 @@ static void ieee80211_work_rx_queued_mgmt(struct ieee80211_local *local,
        case WORK_DONE_REQUEUE:
                synchronize_rcu();
                wk->started = false; /* restart */
-                mutex_lock(&local->work_mtx);
+                mutex_lock(&local->mtx);
                list_add_tail(&wk->list, &local->work_list);
-                mutex_unlock(&local->work_mtx);
+                mutex_unlock(&local->mtx);
        }
 out:
@@ -888,9 +888,9 @@ static void ieee80211_work_work(struct work_struct *work)
        while ((skb = skb_dequeue(&local->work_skb_queue)))
                ieee80211_work_rx_queued_mgmt(local, skb);
-        ieee80211_recalc_idle(local);
+        mutex_lock(&local->mtx);
-        mutex_lock(&local->work_mtx);
+        ieee80211_recalc_idle(local);
        list_for_each_entry_safe(wk, tmp, &local->work_list, list) {
                bool started = wk->started;
@@ -995,20 +995,16 @@ static void ieee80211_work_work(struct work_struct *work)
                run_again(local, jiffies + HZ/2);
        }
-        mutex_lock(&local->scan_mtx);
        if (list_empty(&local->work_list) && local->scan_req &&
            !local->scanning)
                ieee80211_queue_delayed_work(&local->hw,
                                             &local->scan_work,
                                             round_jiffies_relative(0));
-        mutex_unlock(&local->scan_mtx);
-        mutex_unlock(&local->work_mtx);
        ieee80211_recalc_idle(local);
+        mutex_unlock(&local->mtx);
        list_for_each_entry_safe(wk, tmp, &free_work, list) {
                wk->done(wk, NULL);
                list_del(&wk->list);
@@ -1035,16 +1031,15 @@ void ieee80211_add_work(struct ieee80211_work *wk)
        wk->started = false;
        local = wk->sdata->local;
-        mutex_lock(&local->work_mtx);
+        mutex_lock(&local->mtx);
        list_add_tail(&wk->list, &local->work_list);
-        mutex_unlock(&local->work_mtx);
+        mutex_unlock(&local->mtx);
        ieee80211_queue_work(&local->hw, &local->work_work);
 }
 void ieee80211_work_init(struct ieee80211_local *local)
 {
-        mutex_init(&local->work_mtx);
        INIT_LIST_HEAD(&local->work_list);
        setup_timer(&local->work_timer, ieee80211_work_timer,
                    (unsigned long)local);
@@ -1057,7 +1052,7 @@ void ieee80211_work_purge(struct ieee80211_sub_if_data *sdata)
        struct ieee80211_local *local = sdata->local;
        struct ieee80211_work *wk;
-        mutex_lock(&local->work_mtx);
+        mutex_lock(&local->mtx);
        list_for_each_entry(wk, &local->work_list, list) {
                if (wk->sdata != sdata)
                        continue;
@@ -1065,19 +1060,19 @@ void ieee80211_work_purge(struct ieee80211_sub_if_data *sdata)
                wk->started = true;
                wk->timeout = jiffies;
        }
-        mutex_unlock(&local->work_mtx);
+        mutex_unlock(&local->mtx);
        /* run cleanups etc. */
        ieee80211_work_work(&local->work_work);
-        mutex_lock(&local->work_mtx);
+        mutex_lock(&local->mtx);
        list_for_each_entry(wk, &local->work_list, list) {
                if (wk->sdata != sdata)
                        continue;
                WARN_ON(1);
                break;
        }
-        mutex_unlock(&local->work_mtx);
+        mutex_unlock(&local->mtx);
 }
 ieee80211_rx_result ieee80211_work_rx_mgmt(struct ieee80211_sub_if_data *sdata,
@@ -1163,7 +1158,7 @@ int ieee80211_wk_cancel_remain_on_channel(struct ieee80211_sub_if_data *sdata,
        struct ieee80211_work *wk, *tmp;
        bool found = false;
-        mutex_lock(&local->work_mtx);
+        mutex_lock(&local->mtx);
        list_for_each_entry_safe(wk, tmp, &local->work_list, list) {
                if ((unsigned long) wk == cookie) {
                        wk->timeout = jiffies;
@@ -1171,7 +1166,7 @@ int ieee80211_wk_cancel_remain_on_channel(struct ieee80211_sub_if_data *sdata,
                        break;
                }
        }
-        mutex_unlock(&local->work_mtx);
+        mutex_unlock(&local->mtx);
        if (!found)
                return -ENOENT;
diff --git a/net/mac80211/wpa.c b/net/mac80211/wpa.c
index 8d59d27d887e..bee230d8fd11 100644
--- a/net/mac80211/wpa.c
+++ b/net/mac80211/wpa.c
@@ -36,8 +36,8 @@ ieee80211_tx_h_michael_mic_add(struct ieee80211_tx_data *tx)
        int tail;
        hdr = (struct ieee80211_hdr *)skb->data;
-        if (!tx->key || tx->key->conf.alg != ALG_TKIP || skb->len < 24 ||
+        if (!tx->key || tx->key->conf.cipher != WLAN_CIPHER_SUITE_TKIP ||
-            !ieee80211_is_data_present(hdr->frame_control))
+            skb->len < 24 || !ieee80211_is_data_present(hdr->frame_control))
                return TX_CONTINUE;
        hdrlen = ieee80211_hdrlen(hdr->frame_control);
@@ -94,7 +94,7 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx)
        if (status->flag & RX_FLAG_MMIC_STRIPPED)
                return RX_CONTINUE;
-        if (!rx->key || rx->key->conf.alg != ALG_TKIP ||
+        if (!rx->key || rx->key->conf.cipher != WLAN_CIPHER_SUITE_TKIP ||
            !ieee80211_has_protected(hdr->frame_control) ||
            !ieee80211_is_data_present(hdr->frame_control))
                return RX_CONTINUE;
@@ -117,7 +117,7 @@ ieee80211_rx_h_michael_mic_verify(struct ieee80211_rx_data *rx)
        key = &rx->key->conf.key[key_offset];
        michael_mic(key, hdr, data, data_len, mic);
        if (memcmp(mic, data + data_len, MICHAEL_MIC_LEN) != 0 || wpa_test) {
-                if (!(rx->flags & IEEE80211_RX_RA_MATCH))
+                if (!(status->rx_flags & IEEE80211_RX_RA_MATCH))
                        return RX_DROP_UNUSABLE;
                mac80211_ev_michael_mic_failure(rx->sdata, rx->key->conf.keyidx,
@@ -221,19 +221,13 @@ ieee80211_crypto_tkip_decrypt(struct ieee80211_rx_data *rx)
        if (!rx->sta || skb->len - hdrlen < 12)
                return RX_DROP_UNUSABLE;
-        if (status->flag & RX_FLAG_DECRYPTED) {
+        /*
-                if (status->flag & RX_FLAG_IV_STRIPPED) {
+         * Let TKIP code verify IV, but skip decryption.
-                        /*
+         * In the case where hardware checks the IV as well,
-                         * Hardware took care of all processing, including
+         * we don't even get here, see ieee80211_rx_h_decrypt()
-                         * replay protection, and stripped the ICV/IV so
+         */
-                         * we cannot do any checks here.
+        if (status->flag & RX_FLAG_DECRYPTED)
-                         */
-                        return RX_CONTINUE;
-                }
-                /* let TKIP code verify IV, but skip decryption */
                hwaccel = 1;
-        }
        res = ieee80211_tkip_decrypt_data(rx->local->wep_rx_tfm,
                                          key, skb->data + hdrlen,
@@ -447,10 +441,6 @@ ieee80211_crypto_ccmp_decrypt(struct ieee80211_rx_data *rx)
        if (!rx->sta || data_len < 0)
                return RX_DROP_UNUSABLE;
-        if ((status->flag & RX_FLAG_DECRYPTED) &&
-            (status->flag & RX_FLAG_IV_STRIPPED))
-                return RX_CONTINUE;
        ccmp_hdr2pn(pn, skb->data + hdrlen);
        queue = ieee80211_is_mgmt(hdr->frame_control) ?
@@ -564,10 +554,6 @@ ieee80211_crypto_aes_cmac_decrypt(struct ieee80211_rx_data *rx)
        if (!ieee80211_is_mgmt(hdr->frame_control))
                return RX_CONTINUE;
-        if ((status->flag & RX_FLAG_DECRYPTED) &&
-            (status->flag & RX_FLAG_IV_STRIPPED))
-                return RX_CONTINUE;
        if (skb->len < 24 + sizeof(*mmie))
                return RX_DROP_UNUSABLE;
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 4c2f89df5cce..0c043b6ce65e 100644
--- a/net/netfilter/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
@@ -40,6 +40,7 @@
 #include <net/udp.h>
 #include <net/icmp.h>                   /* for icmp_send */
 #include <net/route.h>
+#include <net/ip6_checksum.h>
 #include <linux/netfilter.h>
 #include <linux/netfilter_ipv4.h>
@@ -637,10 +638,12 @@ void ip_vs_nat_icmp_v6(struct sk_buff *skb, struct ip_vs_protocol *pp,
        }
        /* And finally the ICMP checksum */
-        icmph->icmp6_cksum = 0;
+        icmph->icmp6_cksum = ~csum_ipv6_magic(&iph->saddr, &iph->daddr,
-        /* TODO IPv6: is this correct for ICMPv6? */
+                                              skb->len - icmp_offset,
-        ip_vs_checksum_complete(skb, icmp_offset);
+                                              IPPROTO_ICMPV6, 0);
-        skb->ip_summed = CHECKSUM_UNNECESSARY;
+        skb->csum_start = skb_network_header(skb) - skb->head + icmp_offset;
+        skb->csum_offset = offsetof(struct icmp6hdr, icmp6_cksum);
+        skb->ip_summed = CHECKSUM_PARTIAL;
        if (inout)
                IP_VS_DBG_PKT(11, pp, skb, (void *)ciph - (void *)iph,
@@ -1381,8 +1384,7 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb,
        if (af == AF_INET && (ip_vs_sync_state & IP_VS_STATE_MASTER) &&
            cp->protocol == IPPROTO_SCTP) {
                if ((cp->state == IP_VS_SCTP_S_ESTABLISHED &&
-                        (atomic_read(&cp->in_pkts) %
+                        (pkts % sysctl_ip_vs_sync_threshold[1]
-                         sysctl_ip_vs_sync_threshold[1]
                         == sysctl_ip_vs_sync_threshold[0])) ||
                                (cp->old_state != cp->state &&
                                 ((cp->state == IP_VS_SCTP_S_CLOSED) ||
@@ -1393,7 +1395,8 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb,
                }
        }
-        if (af == AF_INET &&
+        /* Keep this block last: TCP and others with pp->num_states <= 1 */
+        else if (af == AF_INET &&
            (ip_vs_sync_state & IP_VS_STATE_MASTER) &&
            (((cp->protocol != IPPROTO_TCP ||
               cp->state == IP_VS_TCP_S_ESTABLISHED) &&
diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
index 0f0c079c422a..ca8ec8c4f311 100644
--- a/net/netfilter/ipvs/ip_vs_ctl.c
+++ b/net/netfilter/ipvs/ip_vs_ctl.c
@@ -61,7 +61,7 @@ static DEFINE_RWLOCK(__ip_vs_svc_lock);
 static DEFINE_RWLOCK(__ip_vs_rs_lock);
 /* lock for state and timeout tables */
-static DEFINE_RWLOCK(__ip_vs_securetcp_lock);
+static DEFINE_SPINLOCK(ip_vs_securetcp_lock);
 /* lock for drop entry handling */
 static DEFINE_SPINLOCK(__ip_vs_dropentry_lock);
@@ -204,7 +204,7 @@ static void update_defense_level(void)
        spin_unlock(&__ip_vs_droppacket_lock);
        /* secure_tcp */
-        write_lock(&__ip_vs_securetcp_lock);
+        spin_lock(&ip_vs_securetcp_lock);
        switch (sysctl_ip_vs_secure_tcp) {
        case 0:
                if (old_secure_tcp >= 2)
@@ -238,7 +238,7 @@ static void update_defense_level(void)
        old_secure_tcp = sysctl_ip_vs_secure_tcp;
        if (to_change >= 0)
                ip_vs_protocol_timeout_change(sysctl_ip_vs_secure_tcp>1);
-        write_unlock(&__ip_vs_securetcp_lock);
+        spin_unlock(&ip_vs_securetcp_lock);
        local_bh_enable();
 }
@@ -843,7 +843,7 @@ ip_vs_new_dest(struct ip_vs_service *svc, struct ip_vs_dest_user_kern *udest,
                        return -EINVAL;
        }
-        dest = kzalloc(sizeof(struct ip_vs_dest), GFP_ATOMIC);
+        dest = kzalloc(sizeof(struct ip_vs_dest), GFP_KERNEL);
        if (dest == NULL) {
                pr_err("%s(): no memory.\n", __func__);
                return -ENOMEM;
@@ -1177,7 +1177,7 @@ ip_vs_add_service(struct ip_vs_service_user_kern *u,
        }
 #endif
-        svc = kzalloc(sizeof(struct ip_vs_service), GFP_ATOMIC);
+        svc = kzalloc(sizeof(struct ip_vs_service), GFP_KERNEL);
        if (svc == NULL) {
                IP_VS_DBG(1, "%s(): no memory\n", __func__);
                ret = -ENOMEM;
@@ -2155,7 +2155,7 @@ do_ip_vs_set_ctl(struct sock *sk, int cmd, void __user *user, unsigned int len)
        if (cmd != IP_VS_SO_SET_ADD
            && (svc == NULL || svc->protocol != usvc.protocol)) {
                ret = -ESRCH;
-                goto out_unlock;
+                goto out_drop_service;
        }
        switch (cmd) {
@@ -2189,6 +2189,7 @@ do_ip_vs_set_ctl(struct sock *sk, int cmd, void __user *user, unsigned int len)
                ret = -EINVAL;
        }
+out_drop_service:
        if (svc)
                ip_vs_service_put(svc);
diff --git a/net/netfilter/ipvs/ip_vs_sched.c b/net/netfilter/ipvs/ip_vs_sched.c
index bbc1ac795952..727e45b66953 100644
--- a/net/netfilter/ipvs/ip_vs_sched.c
+++ b/net/netfilter/ipvs/ip_vs_sched.c
@@ -35,7 +35,7 @@
 static LIST_HEAD(ip_vs_schedulers);
 /* lock for service table */
-static DEFINE_RWLOCK(__ip_vs_sched_lock);
+static DEFINE_SPINLOCK(ip_vs_sched_lock);
 /*
@@ -108,7 +108,7 @@ static struct ip_vs_scheduler *ip_vs_sched_getbyname(const char *sched_name)
        IP_VS_DBG(2, "%s(): sched_name \"%s\"\n", __func__, sched_name);
-        read_lock_bh(&__ip_vs_sched_lock);
+        spin_lock_bh(&ip_vs_sched_lock);
        list_for_each_entry(sched, &ip_vs_schedulers, n_list) {
                /*
@@ -122,14 +122,14 @@ static struct ip_vs_scheduler *ip_vs_sched_getbyname(const char *sched_name)
                }
                if (strcmp(sched_name, sched->name)==0) {
                        /* HIT */
-                        read_unlock_bh(&__ip_vs_sched_lock);
+                        spin_unlock_bh(&ip_vs_sched_lock);
                        return sched;
                }
                if (sched->module)
                        module_put(sched->module);
        }
-        read_unlock_bh(&__ip_vs_sched_lock);
+        spin_unlock_bh(&ip_vs_sched_lock);
        return NULL;
 }
@@ -184,10 +184,10 @@ int register_ip_vs_scheduler(struct ip_vs_scheduler *scheduler)
        /* increase the module use count */
        ip_vs_use_count_inc();
-        write_lock_bh(&__ip_vs_sched_lock);
+        spin_lock_bh(&ip_vs_sched_lock);
        if (!list_empty(&scheduler->n_list)) {
-                write_unlock_bh(&__ip_vs_sched_lock);
+                spin_unlock_bh(&ip_vs_sched_lock);
                ip_vs_use_count_dec();
                pr_err("%s(): [%s] scheduler already linked\n",
                       __func__, scheduler->name);
@@ -200,7 +200,7 @@ int register_ip_vs_scheduler(struct ip_vs_scheduler *scheduler)
         */
        list_for_each_entry(sched, &ip_vs_schedulers, n_list) {
                if (strcmp(scheduler->name, sched->name) == 0) {
-                        write_unlock_bh(&__ip_vs_sched_lock);
+                        spin_unlock_bh(&ip_vs_sched_lock);
                        ip_vs_use_count_dec();
                        pr_err("%s(): [%s] scheduler already existed "
                               "in the system\n", __func__, scheduler->name);
@@ -211,7 +211,7 @@ int register_ip_vs_scheduler(struct ip_vs_scheduler *scheduler)
         *      Add it into the d-linked scheduler list
         */
        list_add(&scheduler->n_list, &ip_vs_schedulers);
-        write_unlock_bh(&__ip_vs_sched_lock);
+        spin_unlock_bh(&ip_vs_sched_lock);
        pr_info("[%s] scheduler registered.\n", scheduler->name);
@@ -229,9 +229,9 @@ int unregister_ip_vs_scheduler(struct ip_vs_scheduler *scheduler)
                return -EINVAL;
        }
-        write_lock_bh(&__ip_vs_sched_lock);
+        spin_lock_bh(&ip_vs_sched_lock);
        if (list_empty(&scheduler->n_list)) {
-                write_unlock_bh(&__ip_vs_sched_lock);
+                spin_unlock_bh(&ip_vs_sched_lock);
                pr_err("%s(): [%s] scheduler is not in the list. failed\n",
                       __func__, scheduler->name);
                return -EINVAL;
@@ -241,7 +241,7 @@ int unregister_ip_vs_scheduler(struct ip_vs_scheduler *scheduler)
         *      Remove it from the d-linked scheduler list
         */
        list_del(&scheduler->n_list);
-        write_unlock_bh(&__ip_vs_sched_lock);
+        spin_unlock_bh(&ip_vs_sched_lock);
        /* decrease the module use count */
        ip_vs_use_count_dec();
diff --git a/net/netfilter/xt_hashlimit.c b/net/netfilter/xt_hashlimit.c
index b46a8390896d..9228ee0dc11a 100644
--- a/net/netfilter/xt_hashlimit.c
+++ b/net/netfilter/xt_hashlimit.c
@@ -448,6 +448,7 @@ hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
 {
        __be16 _ports[2], *ports;
        u8 nexthdr;
+        int poff;
        memset(dst, 0, sizeof(*dst));
@@ -492,19 +493,13 @@ hashlimit_init_dst(const struct xt_hashlimit_htable *hinfo,
                return 0;
        }
-        switch (nexthdr) {
+        poff = proto_ports_offset(nexthdr);
-        case IPPROTO_TCP:
+        if (poff >= 0) {
-        case IPPROTO_UDP:
+                ports = skb_header_pointer(skb, protoff + poff, sizeof(_ports),
-        case IPPROTO_UDPLITE:
-        case IPPROTO_SCTP:
-        case IPPROTO_DCCP:
-                ports = skb_header_pointer(skb, protoff, sizeof(_ports),
                                           &_ports);
-                break;
+        } else {
-        default:
                _ports[0] = _ports[1] = 0;
                ports = _ports;
-                break;
        }
        if (!ports)
                return -1;
diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c
index 9a17f28b1253..3616f27b9d46 100644
--- a/net/packet/af_packet.c
+++ b/net/packet/af_packet.c
@@ -488,7 +488,7 @@ retry:
        skb->dev = dev;
        skb->priority = sk->sk_priority;
        skb->mark = sk->sk_mark;
-        err = sock_tx_timestamp(msg, sk, skb_tx(skb));
+        err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
        if (err < 0)
                goto out_unlock;
@@ -1209,7 +1209,7 @@ static int packet_snd(struct socket *sock,
        err = skb_copy_datagram_from_iovec(skb, offset, msg->msg_iov, 0, len);
        if (err)
                goto out_free;
-        err = sock_tx_timestamp(msg, sk, skb_tx(skb));
+        err = sock_tx_timestamp(sk, &skb_shinfo(skb)->tx_flags);
        if (err < 0)
                goto out_free;
diff --git a/net/phonet/Kconfig b/net/phonet/Kconfig
index 6ec7d55b1769..901956ada9c8 100644
--- a/net/phonet/Kconfig
+++ b/net/phonet/Kconfig
@@ -14,3 +14,14 @@ config PHONET
          To compile this driver as a module, choose M here: the module
          will be called phonet. If unsure, say N.
+config PHONET_PIPECTRLR
+        bool "Phonet Pipe Controller"
+        depends on PHONET
+        default N
+        help
+          The Pipe Controller implementation in Phonet stack to support Pipe
+          data with Nokia Slim modems like WG2.5 used on ST-Ericsson U8500
+          platform.
+          If unsure, say N.
diff --git a/net/phonet/af_phonet.c b/net/phonet/af_phonet.c
index 73aee7f2fcdc..fd95beb72f5d 100644
--- a/net/phonet/af_phonet.c
+++ b/net/phonet/af_phonet.c
@@ -251,6 +251,16 @@ int pn_skb_send(struct sock *sk, struct sk_buff *skb,
        else if (phonet_address_lookup(net, daddr) == 0) {
                dev = phonet_device_get(net);
                skb->pkt_type = PACKET_LOOPBACK;
+        } else if (pn_sockaddr_get_object(target) == 0) {
+                /* Resource routing (small race until phonet_rcv()) */
+                struct sock *sk = pn_find_sock_by_res(net,
+                                                        target->spn_resource);
+                if (sk) {
+                        sock_put(sk);
+                        dev = phonet_device_get(net);
+                        skb->pkt_type = PACKET_LOOPBACK;
+                } else
+                        dev = phonet_route_output(net, daddr);
        } else
                dev = phonet_route_output(net, daddr);
@@ -383,6 +393,13 @@ static int phonet_rcv(struct sk_buff *skb, struct net_device *dev,
                goto out;
        }
+        /* resource routing */
+        if (pn_sockaddr_get_object(&sa) == 0) {
+                struct sock *sk = pn_find_sock_by_res(net, sa.spn_resource);
+                if (sk)
+                        return sk_receive_skb(sk, skb, 0);
+        }
        /* check if we are the destination */
        if (phonet_address_lookup(net, pn_sockaddr_get_addr(&sa)) == 0) {
                /* Phonet packet input */
diff --git a/net/phonet/datagram.c b/net/phonet/datagram.c
index 1bd38db4fe1e..2f032381bd45 100644
--- a/net/phonet/datagram.c
+++ b/net/phonet/datagram.c
@@ -52,6 +52,19 @@ static int pn_ioctl(struct sock *sk, int cmd, unsigned long arg)
                answ = skb ? skb->len : 0;
                release_sock(sk);
                return put_user(answ, (int __user *)arg);
+        case SIOCPNADDRESOURCE:
+        case SIOCPNDELRESOURCE: {
+                        u32 res;
+                        if (get_user(res, (u32 __user *)arg))
+                                return -EFAULT;
+                        if (res >= 256)
+                                return -EINVAL;
+                        if (cmd == SIOCPNADDRESOURCE)
+                                return pn_sock_bind_res(sk, res);
+                        else
+                                return pn_sock_unbind_res(sk, res);
+                }
        }
        return -ENOIOCTLCMD;
diff --git a/net/phonet/pep.c b/net/phonet/pep.c
index 15003021f4f0..aa3d8700d213 100644
--- a/net/phonet/pep.c
+++ b/net/phonet/pep.c
@@ -88,6 +88,15 @@ static int pep_reply(struct sock *sk, struct sk_buff *oskb,
        const struct pnpipehdr *oph = pnp_hdr(oskb);
        struct pnpipehdr *ph;
        struct sk_buff *skb;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        const struct phonethdr *hdr = pn_hdr(oskb);
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = hdr->pn_sdev,
+                .spn_obj = hdr->pn_sobj,
+        };
+#endif
        skb = alloc_skb(MAX_PNPIPE_HEADER + len, priority);
        if (!skb)
@@ -105,10 +114,271 @@ static int pep_reply(struct sock *sk, struct sk_buff *oskb,
        ph->pipe_handle = oph->pipe_handle;
        ph->error_code = code;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        return pn_skb_send(sk, skb, &spn);
+#else
        return pn_skb_send(sk, skb, &pipe_srv);
+#endif
 }
 #define PAD 0x00
+#ifdef CONFIG_PHONET_PIPECTRLR
+static u8 pipe_negotiate_fc(u8 *host_fc, u8 *remote_fc, int len)
+{
+        int i, j;
+        u8 base_fc, final_fc;
+        for (i = 0; i < len; i++) {
+                base_fc = host_fc[i];
+                for (j = 0; j < len; j++) {
+                        if (remote_fc[j] == base_fc) {
+                                final_fc = base_fc;
+                                goto done;
+                        }
+                }
+        }
+        return -EINVAL;
+done:
+        return final_fc;
+}
+static int pipe_get_flow_info(struct sock *sk, struct sk_buff *skb,
+                u8 *pref_rx_fc, u8 *req_tx_fc)
+{
+        struct pnpipehdr *hdr;
+        u8 n_sb;
+        if (!pskb_may_pull(skb, sizeof(*hdr) + 4))
+                return -EINVAL;
+        hdr = pnp_hdr(skb);
+        n_sb = hdr->data[4];
+        __skb_pull(skb, sizeof(*hdr) + 4);
+        while (n_sb > 0) {
+                u8 type, buf[3], len = sizeof(buf);
+                u8 *data = pep_get_sb(skb, &type, &len, buf);
+                if (data == NULL)
+                        return -EINVAL;
+                switch (type) {
+                case PN_PIPE_SB_REQUIRED_FC_TX:
+                        if (len < 3 || (data[2] | data[3] | data[4]) > 3)
+                                break;
+                        req_tx_fc[0] = data[2];
+                        req_tx_fc[1] = data[3];
+                        req_tx_fc[2] = data[4];
+                        break;
+                case PN_PIPE_SB_PREFERRED_FC_RX:
+                        if (len < 3 || (data[2] | data[3] | data[4]) > 3)
+                                break;
+                        pref_rx_fc[0] = data[2];
+                        pref_rx_fc[1] = data[3];
+                        pref_rx_fc[2] = data[4];
+                        break;
+                }
+                n_sb--;
+        }
+        return 0;
+}
+static int pipe_handler_send_req(struct sock *sk, u16 dobj, u8 utid,
+                u8 msg_id, u8 p_handle, gfp_t priority)
+{
+        int len;
+        struct pnpipehdr *ph;
+        struct sk_buff *skb;
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = pn_dev(dobj),
+                .spn_obj = pn_obj(dobj),
+        };
+        static const u8 data[4] = {
+                PAD, PAD, PAD, PAD,
+        };
+        switch (msg_id) {
+        case PNS_PEP_CONNECT_REQ:
+                len = sizeof(data);
+                break;
+        case PNS_PEP_DISCONNECT_REQ:
+        case PNS_PEP_ENABLE_REQ:
+        case PNS_PEP_DISABLE_REQ:
+                len = 0;
+                break;
+        default:
+                return -EINVAL;
+        }
+        skb = alloc_skb(MAX_PNPIPE_HEADER + len, priority);
+        if (!skb)
+                return -ENOMEM;
+        skb_set_owner_w(skb, sk);
+        skb_reserve(skb, MAX_PNPIPE_HEADER);
+        if (len) {
+                __skb_put(skb, len);
+                skb_copy_to_linear_data(skb, data, len);
+        }
+        __skb_push(skb, sizeof(*ph));
+        skb_reset_transport_header(skb);
+        ph = pnp_hdr(skb);
+        ph->utid = utid;
+        ph->message_id = msg_id;
+        ph->pipe_handle = p_handle;
+        ph->error_code = PN_PIPE_NO_ERROR;
+        return pn_skb_send(sk, skb, &spn);
+}
+static int pipe_handler_send_created_ind(struct sock *sk, u16 dobj,
+                u8 utid, u8 p_handle, u8 msg_id, u8 tx_fc, u8 rx_fc)
+{
+        int err_code;
+        struct pnpipehdr *ph;
+        struct sk_buff *skb;
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = pn_dev(dobj),
+                .spn_obj = pn_obj(dobj),
+        };
+        static u8 data[4] = {
+                0x03, 0x04,
+        };
+        data[2] = tx_fc;
+        data[3] = rx_fc;
+        /*
+         * actually, below is number of sub-blocks and not error code.
+         * Pipe_created_ind message format does not have any
+         * error code field. However, the Phonet stack will always send
+         * an error code as part of pnpipehdr. So, use that err_code to
+         * specify the number of sub-blocks.
+         */
+        err_code = 0x01;
+        skb = alloc_skb(MAX_PNPIPE_HEADER + sizeof(data), GFP_ATOMIC);
+        if (!skb)
+                return -ENOMEM;
+        skb_set_owner_w(skb, sk);
+        skb_reserve(skb, MAX_PNPIPE_HEADER);
+        __skb_put(skb, sizeof(data));
+        skb_copy_to_linear_data(skb, data, sizeof(data));
+        __skb_push(skb, sizeof(*ph));
+        skb_reset_transport_header(skb);
+        ph = pnp_hdr(skb);
+        ph->utid = utid;
+        ph->message_id = msg_id;
+        ph->pipe_handle = p_handle;
+        ph->error_code = err_code;
+        return pn_skb_send(sk, skb, &spn);
+}
+static int pipe_handler_send_ind(struct sock *sk, u16 dobj, u8 utid,
+                u8 p_handle, u8 msg_id)
+{
+        int err_code;
+        struct pnpipehdr *ph;
+        struct sk_buff *skb;
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = pn_dev(dobj),
+                .spn_obj = pn_obj(dobj),
+        };
+        /*
+         * actually, below is a filler.
+         * Pipe_enabled/disabled_ind message format does not have any
+         * error code field. However, the Phonet stack will always send
+         * an error code as part of pnpipehdr. So, use that err_code to
+         * specify the filler value.
+         */
+        err_code = 0x0;
+        skb = alloc_skb(MAX_PNPIPE_HEADER, GFP_ATOMIC);
+        if (!skb)
+                return -ENOMEM;
+        skb_set_owner_w(skb, sk);
+        skb_reserve(skb, MAX_PNPIPE_HEADER);
+        __skb_push(skb, sizeof(*ph));
+        skb_reset_transport_header(skb);
+        ph = pnp_hdr(skb);
+        ph->utid = utid;
+        ph->message_id = msg_id;
+        ph->pipe_handle = p_handle;
+        ph->error_code = err_code;
+        return pn_skb_send(sk, skb, &spn);
+}
+static int pipe_handler_enable_pipe(struct sock *sk, int cmd)
+{
+        int ret;
+        struct pep_sock *pn = pep_sk(sk);
+        switch (cmd) {
+        case PNPIPE_ENABLE:
+                ret = pipe_handler_send_req(sk, pn->pn_sk.sobject,
+                                PNS_PIPE_ENABLE_UTID, PNS_PEP_ENABLE_REQ,
+                                pn->pipe_handle, GFP_ATOMIC);
+                break;
+        case PNPIPE_DISABLE:
+                ret = pipe_handler_send_req(sk, pn->pn_sk.sobject,
+                                PNS_PIPE_DISABLE_UTID, PNS_PEP_DISABLE_REQ,
+                                pn->pipe_handle, GFP_ATOMIC);
+                break;
+        default:
+                ret = -EINVAL;
+        }
+        return ret;
+}
+static int pipe_handler_create_pipe(struct sock *sk, int pipe_handle, int cmd)
+{
+        int ret;
+        struct pep_sock *pn = pep_sk(sk);
+        switch (cmd) {
+        case PNPIPE_CREATE:
+                ret = pipe_handler_send_req(sk, pn->pn_sk.sobject,
+                                PNS_PEP_CONNECT_UTID, PNS_PEP_CONNECT_REQ,
+                                pipe_handle, GFP_ATOMIC);
+                break;
+        case PNPIPE_DESTROY:
+                ret = pipe_handler_send_req(sk, pn->remote_pep,
+                                PNS_PEP_DISCONNECT_UTID,
+                                PNS_PEP_DISCONNECT_REQ,
+                                pn->pipe_handle, GFP_ATOMIC);
+                break;
+        default:
+                ret = -EINVAL;
+        }
+        return ret;
+}
+#endif
 static int pep_accept_conn(struct sock *sk, struct sk_buff *skb)
 {
        static const u8 data[20] = {
@@ -173,6 +443,14 @@ static int pipe_snd_status(struct sock *sk, u8 type, u8 status, gfp_t priority)
        struct pep_sock *pn = pep_sk(sk);
        struct pnpipehdr *ph;
        struct sk_buff *skb;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = pn_dev(pn->remote_pep),
+                .spn_obj = pn_obj(pn->remote_pep),
+        };
+#endif
        skb = alloc_skb(MAX_PNPIPE_HEADER + 4, priority);
        if (!skb)
@@ -192,7 +470,11 @@ static int pipe_snd_status(struct sock *sk, u8 type, u8 status, gfp_t priority)
        ph->data[3] = PAD;
        ph->data[4] = status;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        return pn_skb_send(sk, skb, &spn);
+#else
        return pn_skb_send(sk, skb, &pipe_srv);
+#endif
 }
 /* Send our RX flow control information to the sender.
@@ -309,6 +591,12 @@ static int pipe_do_rcv(struct sock *sk, struct sk_buff *skb)
        struct pnpipehdr *hdr = pnp_hdr(skb);
        struct sk_buff_head *queue;
        int err = 0;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        struct phonethdr *ph = pn_hdr(skb);
+        static u8 host_pref_rx_fc[3], host_req_tx_fc[3];
+        u8 remote_pref_rx_fc[3], remote_req_tx_fc[3];
+        u8 negotiated_rx_fc, negotiated_tx_fc;
+#endif
        BUG_ON(sk->sk_state == TCP_CLOSE_WAIT);
@@ -317,6 +605,40 @@ static int pipe_do_rcv(struct sock *sk, struct sk_buff *skb)
                pep_reject_conn(sk, skb, PN_PIPE_ERR_PEP_IN_USE);
                break;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNS_PEP_CONNECT_RESP:
+                if ((ph->pn_sdev == pn_dev(pn->remote_pep)) &&
+                                (ph->pn_sobj == pn_obj(pn->remote_pep))) {
+                        pipe_get_flow_info(sk, skb, remote_pref_rx_fc,
+                                        remote_req_tx_fc);
+                         negotiated_tx_fc = pipe_negotiate_fc(remote_req_tx_fc,
+                                         host_pref_rx_fc,
+                                         sizeof(host_pref_rx_fc));
+                         negotiated_rx_fc = pipe_negotiate_fc(host_req_tx_fc,
+                                         remote_pref_rx_fc,
+                                         sizeof(host_pref_rx_fc));
+                        pn->pipe_state = PIPE_DISABLED;
+                        pipe_handler_send_created_ind(sk, pn->remote_pep,
+                                        PNS_PIPE_CREATED_IND_UTID,
+                                        pn->pipe_handle, PNS_PIPE_CREATED_IND,
+                                        negotiated_tx_fc, negotiated_rx_fc);
+                        pipe_handler_send_created_ind(sk, pn->pn_sk.sobject,
+                                        PNS_PIPE_CREATED_IND_UTID,
+                                        pn->pipe_handle, PNS_PIPE_CREATED_IND,
+                                        negotiated_tx_fc, negotiated_rx_fc);
+                } else {
+                        pipe_handler_send_req(sk, pn->remote_pep,
+                                        PNS_PEP_CONNECT_UTID,
+                                        PNS_PEP_CONNECT_REQ, pn->pipe_handle,
+                                        GFP_ATOMIC);
+                        pipe_get_flow_info(sk, skb, host_pref_rx_fc,
+                                        host_req_tx_fc);
+                }
+                break;
+#endif
        case PNS_PEP_DISCONNECT_REQ:
                pep_reply(sk, skb, PN_PIPE_NO_ERROR, NULL, 0, GFP_ATOMIC);
                sk->sk_state = TCP_CLOSE_WAIT;
@@ -324,11 +646,41 @@ static int pipe_do_rcv(struct sock *sk, struct sk_buff *skb)
                        sk->sk_state_change(sk);
                break;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNS_PEP_DISCONNECT_RESP:
+                pn->pipe_state = PIPE_IDLE;
+                pipe_handler_send_req(sk, pn->pn_sk.sobject,
+                                PNS_PEP_DISCONNECT_UTID,
+                                PNS_PEP_DISCONNECT_REQ, pn->pipe_handle,
+                                GFP_KERNEL);
+                break;
+#endif
        case PNS_PEP_ENABLE_REQ:
                /* Wait for PNS_PIPE_(ENABLED|REDIRECTED)_IND */
                pep_reply(sk, skb, PN_PIPE_NO_ERROR, NULL, 0, GFP_ATOMIC);
                break;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNS_PEP_ENABLE_RESP:
+                if ((ph->pn_sdev == pn_dev(pn->remote_pep)) &&
+                                (ph->pn_sobj == pn_obj(pn->remote_pep))) {
+                        pn->pipe_state = PIPE_ENABLED;
+                        pipe_handler_send_ind(sk, pn->remote_pep,
+                                        PNS_PIPE_ENABLED_IND_UTID,
+                                        pn->pipe_handle, PNS_PIPE_ENABLED_IND);
+                        pipe_handler_send_ind(sk, pn->pn_sk.sobject,
+                                        PNS_PIPE_ENABLED_IND_UTID,
+                                        pn->pipe_handle, PNS_PIPE_ENABLED_IND);
+                } else
+                        pipe_handler_send_req(sk, pn->remote_pep,
+                                        PNS_PIPE_ENABLE_UTID,
+                                        PNS_PEP_ENABLE_REQ, pn->pipe_handle,
+                                        GFP_KERNEL);
+                break;
+#endif
        case PNS_PEP_RESET_REQ:
                switch (hdr->state_after_reset) {
                case PN_PIPE_DISABLE:
@@ -347,6 +699,27 @@ static int pipe_do_rcv(struct sock *sk, struct sk_buff *skb)
                pep_reply(sk, skb, PN_PIPE_NO_ERROR, NULL, 0, GFP_ATOMIC);
                break;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNS_PEP_DISABLE_RESP:
+                if ((ph->pn_sdev == pn_dev(pn->remote_pep)) &&
+                                (ph->pn_sobj == pn_obj(pn->remote_pep))) {
+                        pn->pipe_state = PIPE_DISABLED;
+                        pipe_handler_send_ind(sk, pn->remote_pep,
+                                        PNS_PIPE_DISABLED_IND_UTID,
+                                        pn->pipe_handle,
+                                        PNS_PIPE_DISABLED_IND);
+                        pipe_handler_send_ind(sk, pn->pn_sk.sobject,
+                                        PNS_PIPE_DISABLED_IND_UTID,
+                                        pn->pipe_handle,
+                                        PNS_PIPE_DISABLED_IND);
+                } else
+                        pipe_handler_send_req(sk, pn->remote_pep,
+                                        PNS_PIPE_DISABLE_UTID,
+                                        PNS_PEP_DISABLE_REQ, pn->pipe_handle,
+                                        GFP_KERNEL);
+                break;
+#endif
        case PNS_PEP_CTRL_REQ:
                if (skb_queue_len(&pn->ctrlreq_queue) >= PNPIPE_CTRLREQ_MAX) {
                        atomic_inc(&sk->sk_drops);
@@ -520,6 +893,9 @@ static int pep_connreq_rcv(struct sock *sk, struct sk_buff *skb)
        newpn->rx_fc = newpn->tx_fc = PN_LEGACY_FLOW_CONTROL;
        newpn->init_enable = enabled;
        newpn->aligned = aligned;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        newpn->remote_pep = pn->remote_pep;
+#endif
        BUG_ON(!skb_queue_empty(&newsk->sk_receive_queue));
        skb_queue_head(&newsk->sk_receive_queue, skb);
@@ -621,6 +997,28 @@ drop:
        return err;
 }
+static int pipe_do_remove(struct sock *sk)
+{
+        struct pep_sock *pn = pep_sk(sk);
+        struct pnpipehdr *ph;
+        struct sk_buff *skb;
+        skb = alloc_skb(MAX_PNPIPE_HEADER, GFP_KERNEL);
+        if (!skb)
+                return -ENOMEM;
+        skb_reserve(skb, MAX_PNPIPE_HEADER);
+        __skb_push(skb, sizeof(*ph));
+        skb_reset_transport_header(skb);
+        ph = pnp_hdr(skb);
+        ph->utid = 0;
+        ph->message_id = PNS_PIPE_REMOVE_REQ;
+        ph->pipe_handle = pn->pipe_handle;
+        ph->data[0] = PAD;
+        return pn_skb_send(sk, skb, &pipe_srv);
+}
 /* associated socket ceases to exist */
 static void pep_sock_close(struct sock *sk, long timeout)
 {
@@ -639,7 +1037,10 @@ static void pep_sock_close(struct sock *sk, long timeout)
                sk_for_each_safe(sknode, p, n, &pn->ackq)
                        sk_del_node_init(sknode);
                sk->sk_state = TCP_CLOSE;
-        }
+        } else if ((1 << sk->sk_state) & (TCPF_SYN_RECV|TCPF_ESTABLISHED))
+                /* Forcefully remove dangling Phonet pipe */
+                pipe_do_remove(sk);
        ifindex = pn->ifindex;
        pn->ifindex = 0;
        release_sock(sk);
@@ -757,6 +1158,10 @@ static int pep_setsockopt(struct sock *sk, int level, int optname,
 {
        struct pep_sock *pn = pep_sk(sk);
        int val = 0, err = 0;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        int remote_pep;
+        int pipe_handle;
+#endif
        if (level != SOL_PNPIPE)
                return -ENOPROTOOPT;
@@ -767,6 +1172,48 @@ static int pep_setsockopt(struct sock *sk, int level, int optname,
        lock_sock(sk);
        switch (optname) {
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNPIPE_CREATE:
+                if (val) {
+                        if (pn->pipe_state > PIPE_IDLE) {
+                                err = -EFAULT;
+                                break;
+                        }
+                        remote_pep = val & 0xFFFF;
+                        pipe_handle =  (val >> 16) & 0xFF;
+                        pn->remote_pep = remote_pep;
+                        err = pipe_handler_create_pipe(sk, pipe_handle,
+                                        PNPIPE_CREATE);
+                        break;
+                }
+        case PNPIPE_ENABLE:
+                if (pn->pipe_state != PIPE_DISABLED) {
+                        err = -EFAULT;
+                        break;
+                }
+                err = pipe_handler_enable_pipe(sk, PNPIPE_ENABLE);
+                break;
+        case PNPIPE_DISABLE:
+                if (pn->pipe_state != PIPE_ENABLED) {
+                        err = -EFAULT;
+                        break;
+                }
+                err = pipe_handler_enable_pipe(sk, PNPIPE_DISABLE);
+                break;
+        case PNPIPE_DESTROY:
+                if (pn->pipe_state < PIPE_DISABLED) {
+                        err = -EFAULT;
+                        break;
+                }
+                err = pipe_handler_create_pipe(sk, 0x0, PNPIPE_DESTROY);
+                break;
+#endif
        case PNPIPE_ENCAP:
                if (val && val != PNPIPE_ENCAP_IP) {
                        err = -EINVAL;
@@ -816,6 +1263,13 @@ static int pep_getsockopt(struct sock *sk, int level, int optname,
        case PNPIPE_ENCAP:
                val = pn->ifindex ? PNPIPE_ENCAP_IP : PNPIPE_ENCAP_NONE;
                break;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        case PNPIPE_INQ:
+                val = pn->pipe_state;
+                break;
+#endif
        case PNPIPE_IFINDEX:
                val = pn->ifindex;
                break;
@@ -835,6 +1289,15 @@ static int pipe_skb_send(struct sock *sk, struct sk_buff *skb)
 {
        struct pep_sock *pn = pep_sk(sk);
        struct pnpipehdr *ph;
+        int err;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        struct sockaddr_pn spn = {
+                .spn_family = AF_PHONET,
+                .spn_resource = 0xD9,
+                .spn_dev = pn_dev(pn->remote_pep),
+                .spn_obj = pn_obj(pn->remote_pep),
+        };
+#endif
        if (pn_flow_safe(pn->tx_fc) &&
            !atomic_add_unless(&pn->tx_credits, -1, 0)) {
@@ -852,8 +1315,16 @@ static int pipe_skb_send(struct sock *sk, struct sk_buff *skb)
        } else
                ph->message_id = PNS_PIPE_DATA;
        ph->pipe_handle = pn->pipe_handle;
+#ifdef CONFIG_PHONET_PIPECTRLR
+        err = pn_skb_send(sk, skb, &spn);
+#else
+        err = pn_skb_send(sk, skb, &pipe_srv);
+#endif
+        if (err && pn_flow_safe(pn->tx_fc))
+                atomic_inc(&pn->tx_credits);
+        return err;
-        return pn_skb_send(sk, skb, &pipe_srv);
 }
 static int pep_sendmsg(struct kiocb *iocb, struct sock *sk,
@@ -873,7 +1344,7 @@ static int pep_sendmsg(struct kiocb *iocb, struct sock *sk,
        skb = sock_alloc_send_skb(sk, MAX_PNPIPE_HEADER + len,
                                        flags & MSG_DONTWAIT, &err);
        if (!skb)
-                return -ENOBUFS;
+                return err;
        skb_reserve(skb, MAX_PHONET_HEADER + 3);
        err = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);
diff --git a/net/phonet/pn_dev.c b/net/phonet/pn_dev.c
index b18e48fae975..947038ddd04c 100644
--- a/net/phonet/pn_dev.c
+++ b/net/phonet/pn_dev.c
@@ -292,8 +292,7 @@ static void phonet_route_autodel(struct net_device *dev)
        if (bitmap_empty(deleted, 64))
                return; /* short-circuit RCU */
        synchronize_rcu();
-        for (i = find_first_bit(deleted, 64); i < 64;
+        for_each_set_bit(i, deleted, 64) {
-                        i = find_next_bit(deleted, 64, i + 1)) {
                rtm_phonet_notify(RTM_DELROUTE, dev, i);
                dev_put(dev);
        }
@@ -374,6 +373,7 @@ int __init phonet_device_init(void)
        if (err)
                return err;
+        proc_net_fops_create(&init_net, "pnresource", 0, &pn_res_seq_fops);
        register_netdevice_notifier(&phonet_device_notifier);
        err = phonet_netlink_register();
        if (err)
@@ -386,6 +386,7 @@ void phonet_device_exit(void)
        rtnl_unregister_all(PF_PHONET);
        unregister_netdevice_notifier(&phonet_device_notifier);
        unregister_pernet_device(&phonet_net_ops);
+        proc_net_remove(&init_net, "pnresource");
 }
 int phonet_route_add(struct net_device *dev, u8 daddr)
diff --git a/net/phonet/socket.c b/net/phonet/socket.c
index 6e9848bf0370..aca8fba099e9 100644
--- a/net/phonet/socket.c
+++ b/net/phonet/socket.c
@@ -158,6 +158,7 @@ void pn_sock_unhash(struct sock *sk)
        spin_lock_bh(&pnsocks.lock);
        sk_del_node_init(sk);
        spin_unlock_bh(&pnsocks.lock);
+        pn_sock_unbind_all_res(sk);
 }
 EXPORT_SYMBOL(pn_sock_unhash);
@@ -281,7 +282,9 @@ static unsigned int pn_socket_poll(struct file *file, struct socket *sock,
        if (!mask && sk->sk_state == TCP_CLOSE_WAIT)
                return POLLHUP;
-        if (sk->sk_state == TCP_ESTABLISHED && atomic_read(&pn->tx_credits))
+        if (sk->sk_state == TCP_ESTABLISHED &&
+                atomic_read(&sk->sk_wmem_alloc) < sk->sk_sndbuf &&
+                atomic_read(&pn->tx_credits))
                mask |= POLLOUT | POLLWRNORM | POLLWRBAND;
        return mask;
@@ -563,3 +566,188 @@ const struct file_operations pn_sock_seq_fops = {
        .release = seq_release_net,
 };
 #endif
+static struct  {
+        struct sock *sk[256];
+} pnres;
+/*
+ * Find and hold socket based on resource.
+ */
+struct sock *pn_find_sock_by_res(struct net *net, u8 res)
+{
+        struct sock *sk;
+        if (!net_eq(net, &init_net))
+                return NULL;
+        rcu_read_lock();
+        sk = rcu_dereference(pnres.sk[res]);
+        if (sk)
+                sock_hold(sk);
+        rcu_read_unlock();
+        return sk;
+}
+static DEFINE_MUTEX(resource_mutex);
+int pn_sock_bind_res(struct sock *sk, u8 res)
+{
+        int ret = -EADDRINUSE;
+        if (!net_eq(sock_net(sk), &init_net))
+                return -ENOIOCTLCMD;
+        if (!capable(CAP_SYS_ADMIN))
+                return -EPERM;
+        if (pn_socket_autobind(sk->sk_socket))
+                return -EAGAIN;
+        mutex_lock(&resource_mutex);
+        if (pnres.sk[res] == NULL) {
+                sock_hold(sk);
+                rcu_assign_pointer(pnres.sk[res], sk);
+                ret = 0;
+        }
+        mutex_unlock(&resource_mutex);
+        return ret;
+}
+int pn_sock_unbind_res(struct sock *sk, u8 res)
+{
+        int ret = -ENOENT;
+        if (!capable(CAP_SYS_ADMIN))
+                return -EPERM;
+        mutex_lock(&resource_mutex);
+        if (pnres.sk[res] == sk) {
+                rcu_assign_pointer(pnres.sk[res], NULL);
+                ret = 0;
+        }
+        mutex_unlock(&resource_mutex);
+        if (ret == 0) {
+                synchronize_rcu();
+                sock_put(sk);
+        }
+        return ret;
+}
+void pn_sock_unbind_all_res(struct sock *sk)
+{
+        unsigned res, match = 0;
+        mutex_lock(&resource_mutex);
+        for (res = 0; res < 256; res++) {
+                if (pnres.sk[res] == sk) {
+                        rcu_assign_pointer(pnres.sk[res], NULL);
+                        match++;
+                }
+        }
+        mutex_unlock(&resource_mutex);
+        if (match == 0)
+                return;
+        synchronize_rcu();
+        while (match > 0) {
+                sock_put(sk);
+                match--;
+        }
+}
+#ifdef CONFIG_PROC_FS
+static struct sock **pn_res_get_idx(struct seq_file *seq, loff_t pos)
+{
+        struct net *net = seq_file_net(seq);
+        unsigned i;
+        if (!net_eq(net, &init_net))
+                return NULL;
+        for (i = 0; i < 256; i++) {
+                if (pnres.sk[i] == NULL)
+                        continue;
+                if (!pos)
+                        return pnres.sk + i;
+                pos--;
+        }
+        return NULL;
+}
+static struct sock **pn_res_get_next(struct seq_file *seq, struct sock **sk)
+{
+        struct net *net = seq_file_net(seq);
+        unsigned i;
+        BUG_ON(!net_eq(net, &init_net));
+        for (i = (sk - pnres.sk) + 1; i < 256; i++)
+                if (pnres.sk[i])
+                        return pnres.sk + i;
+        return NULL;
+}
+static void *pn_res_seq_start(struct seq_file *seq, loff_t *pos)
+        __acquires(resource_mutex)
+{
+        mutex_lock(&resource_mutex);
+        return *pos ? pn_res_get_idx(seq, *pos - 1) : SEQ_START_TOKEN;
+}
+static void *pn_res_seq_next(struct seq_file *seq, void *v, loff_t *pos)
+{
+        struct sock **sk;
+        if (v == SEQ_START_TOKEN)
+                sk = pn_res_get_idx(seq, 0);
+        else
+                sk = pn_res_get_next(seq, v);
+        (*pos)++;
+        return sk;
+}
+static void pn_res_seq_stop(struct seq_file *seq, void *v)
+        __releases(resource_mutex)
+{
+        mutex_unlock(&resource_mutex);
+}
+static int pn_res_seq_show(struct seq_file *seq, void *v)
+{
+        int len;
+        if (v == SEQ_START_TOKEN)
+                seq_printf(seq, "%s%n", "rs   uid inode", &len);
+        else {
+                struct sock **psk = v;
+                struct sock *sk = *psk;
+                seq_printf(seq, "%02X %5d %lu%n",
+                           (int) (psk - pnres.sk), sock_i_uid(sk),
+                           sock_i_ino(sk), &len);
+        }
+        seq_printf(seq, "%*s\n", 63 - len, "");
+        return 0;
+}
+static const struct seq_operations pn_res_seq_ops = {
+        .start = pn_res_seq_start,
+        .next = pn_res_seq_next,
+        .stop = pn_res_seq_stop,
+        .show = pn_res_seq_show,
+};
+static int pn_res_open(struct inode *inode, struct file *file)
+{
+        return seq_open_net(inode, file, &pn_res_seq_ops,
+                                sizeof(struct seq_net_private));
+}
+const struct file_operations pn_res_seq_fops = {
+        .owner = THIS_MODULE,
+        .open = pn_res_open,
+        .read = seq_read,
+        .llseek = seq_lseek,
+        .release = seq_release_net,
+};
+#endif
diff --git a/net/rds/af_rds.c b/net/rds/af_rds.c
index aebfecbdb841..bb6ad81b671d 100644
--- a/net/rds/af_rds.c
+++ b/net/rds/af_rds.c
@@ -39,7 +39,15 @@
 #include <net/sock.h>
 #include "rds.h"
-#include "rdma.h"
+char *rds_str_array(char **array, size_t elements, size_t index)
+{
+        if ((index < elements) && array[index])
+                return array[index];
+        else
+                return "unknown";
+}
+EXPORT_SYMBOL(rds_str_array);
 /* this is just used for stats gathering :/ */
 static DEFINE_SPINLOCK(rds_sock_lock);
@@ -62,7 +70,7 @@ static int rds_release(struct socket *sock)
        struct rds_sock *rs;
        unsigned long flags;
-        if (sk == NULL)
+        if (!sk)
                goto out;
        rs = rds_sk_to_rs(sk);
@@ -73,7 +81,15 @@ static int rds_release(struct socket *sock)
         * with the socket. */
        rds_clear_recv_queue(rs);
        rds_cong_remove_socket(rs);
+        /*
+         * the binding lookup hash uses rcu, we need to
+         * make sure we sychronize_rcu before we free our
+         * entry
+         */
        rds_remove_bound(rs);
+        synchronize_rcu();
        rds_send_drop_to(rs, NULL);
        rds_rdma_drop_keys(rs);
        rds_notify_queue_get(rs, NULL);
@@ -83,6 +99,8 @@ static int rds_release(struct socket *sock)
        rds_sock_count--;
        spin_unlock_irqrestore(&rds_sock_lock, flags);
+        rds_trans_put(rs->rs_transport);
        sock->sk = NULL;
        sock_put(sk);
 out:
@@ -514,7 +532,7 @@ out:
        spin_unlock_irqrestore(&rds_sock_lock, flags);
 }
-static void __exit rds_exit(void)
+static void rds_exit(void)
 {
        sock_unregister(rds_family_ops.family);
        proto_unregister(&rds_proto);
@@ -529,7 +547,7 @@ static void __exit rds_exit(void)
 }
 module_exit(rds_exit);
-static int __init rds_init(void)
+static int rds_init(void)
 {
        int ret;
diff --git a/net/rds/bind.c b/net/rds/bind.c
index 5d95fc007f1a..2f6b3fcc79f8 100644
--- a/net/rds/bind.c
+++ b/net/rds/bind.c
@@ -34,45 +34,52 @@
 #include <net/sock.h>
 #include <linux/in.h>
 #include <linux/if_arp.h>
+#include <linux/jhash.h>
 #include "rds.h"
-/*
+#define BIND_HASH_SIZE 1024
- * XXX this probably still needs more work.. no INADDR_ANY, and rbtrees aren't
+static struct hlist_head bind_hash_table[BIND_HASH_SIZE];
- * particularly zippy.
- *
- * This is now called for every incoming frame so we arguably care much more
- * about it than we used to.
- */
 static DEFINE_SPINLOCK(rds_bind_lock);
-static struct rb_root rds_bind_tree = RB_ROOT;
-static struct rds_sock *rds_bind_tree_walk(__be32 addr, __be16 port,
+static struct hlist_head *hash_to_bucket(__be32 addr, __be16 port)
-                                           struct rds_sock *insert)
+{
+        return bind_hash_table + (jhash_2words((u32)addr, (u32)port, 0) &
+                                  (BIND_HASH_SIZE - 1));
+}
+static struct rds_sock *rds_bind_lookup(__be32 addr, __be16 port,
+                                        struct rds_sock *insert)
 {
-        struct rb_node **p = &rds_bind_tree.rb_node;
-        struct rb_node *parent = NULL;
        struct rds_sock *rs;
+        struct hlist_node *node;
+        struct hlist_head *head = hash_to_bucket(addr, port);
        u64 cmp;
        u64 needle = ((u64)be32_to_cpu(addr) << 32) | be16_to_cpu(port);
-        while (*p) {
+        rcu_read_lock();
-                parent = *p;
+        hlist_for_each_entry_rcu(rs, node, head, rs_bound_node) {
-                rs = rb_entry(parent, struct rds_sock, rs_bound_node);
                cmp = ((u64)be32_to_cpu(rs->rs_bound_addr) << 32) |
                      be16_to_cpu(rs->rs_bound_port);
-                if (needle < cmp)
+                if (cmp == needle) {
-                        p = &(*p)->rb_left;
+                        rcu_read_unlock();
-                else if (needle > cmp)
-                        p = &(*p)->rb_right;
-                else
                        return rs;
+                }
        }
+        rcu_read_unlock();
        if (insert) {
-                rb_link_node(&insert->rs_bound_node, parent, p);
+                /*
-                rb_insert_color(&insert->rs_bound_node, &rds_bind_tree);
+                 * make sure our addr and port are set before
+                 * we are added to the list, other people
+                 * in rcu will find us as soon as the
+                 * hlist_add_head_rcu is done
+                 */
+                insert->rs_bound_addr = addr;
+                insert->rs_bound_port = port;
+                rds_sock_addref(insert);
+                hlist_add_head_rcu(&insert->rs_bound_node, head);
        }
        return NULL;
 }
@@ -86,15 +93,13 @@ static struct rds_sock *rds_bind_tree_walk(__be32 addr, __be16 port,
 struct rds_sock *rds_find_bound(__be32 addr, __be16 port)
 {
        struct rds_sock *rs;
-        unsigned long flags;
-        spin_lock_irqsave(&rds_bind_lock, flags);
+        rs = rds_bind_lookup(addr, port, NULL);
-        rs = rds_bind_tree_walk(addr, port, NULL);
        if (rs && !sock_flag(rds_rs_to_sk(rs), SOCK_DEAD))
                rds_sock_addref(rs);
        else
                rs = NULL;
-        spin_unlock_irqrestore(&rds_bind_lock, flags);
        rdsdebug("returning rs %p for %pI4:%u\n", rs, &addr,
                ntohs(port));
@@ -121,22 +126,15 @@ static int rds_add_bound(struct rds_sock *rs, __be32 addr, __be16 *port)
        do {
                if (rover == 0)
                        rover++;
-                if (rds_bind_tree_walk(addr, cpu_to_be16(rover), rs) == NULL) {
+                if (!rds_bind_lookup(addr, cpu_to_be16(rover), rs)) {
-                        *port = cpu_to_be16(rover);
+                        *port = rs->rs_bound_port;
                        ret = 0;
+                        rdsdebug("rs %p binding to %pI4:%d\n",
+                          rs, &addr, (int)ntohs(*port));
                        break;
                }
        } while (rover++ != last);
-        if (ret == 0)  {
-                rs->rs_bound_addr = addr;
-                rs->rs_bound_port = *port;
-                rds_sock_addref(rs);
-                rdsdebug("rs %p binding to %pI4:%d\n",
-                  rs, &addr, (int)ntohs(*port));
-        }
        spin_unlock_irqrestore(&rds_bind_lock, flags);
        return ret;
@@ -153,7 +151,7 @@ void rds_remove_bound(struct rds_sock *rs)
                  rs, &rs->rs_bound_addr,
                  ntohs(rs->rs_bound_port));
-                rb_erase(&rs->rs_bound_node, &rds_bind_tree);
+                hlist_del_init_rcu(&rs->rs_bound_node);
                rds_sock_put(rs);
                rs->rs_bound_addr = 0;
        }
@@ -184,7 +182,7 @@ int rds_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
                goto out;
        trans = rds_trans_get_preferred(sin->sin_addr.s_addr);
-        if (trans == NULL) {
+        if (!trans) {
                ret = -EADDRNOTAVAIL;
                rds_remove_bound(rs);
                if (printk_ratelimit())
@@ -198,5 +196,9 @@ int rds_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
 out:
        release_sock(sk);
+        /* we might have called rds_remove_bound on error */
+        if (ret)
+                synchronize_rcu();
        return ret;
 }
diff --git a/net/rds/cong.c b/net/rds/cong.c
index 0871a29f0780..75ea686f27d5 100644
--- a/net/rds/cong.c
+++ b/net/rds/cong.c
@@ -141,7 +141,7 @@ static struct rds_cong_map *rds_cong_from_addr(__be32 addr)
        unsigned long flags;
        map = kzalloc(sizeof(struct rds_cong_map), GFP_KERNEL);
-        if (map == NULL)
+        if (!map)
                return NULL;
        map->m_addr = addr;
@@ -159,7 +159,7 @@ static struct rds_cong_map *rds_cong_from_addr(__be32 addr)
        ret = rds_cong_tree_walk(addr, map);
        spin_unlock_irqrestore(&rds_cong_lock, flags);
-        if (ret == NULL) {
+        if (!ret) {
                ret = map;
                map = NULL;
        }
@@ -205,7 +205,7 @@ int rds_cong_get_maps(struct rds_connection *conn)
        conn->c_lcong = rds_cong_from_addr(conn->c_laddr);
        conn->c_fcong = rds_cong_from_addr(conn->c_faddr);
-        if (conn->c_lcong == NULL || conn->c_fcong == NULL)
+        if (!(conn->c_lcong && conn->c_fcong))
                return -ENOMEM;
        return 0;
@@ -221,7 +221,7 @@ void rds_cong_queue_updates(struct rds_cong_map *map)
        list_for_each_entry(conn, &map->m_conn_list, c_map_item) {
                if (!test_and_set_bit(0, &conn->c_map_queued)) {
                        rds_stats_inc(s_cong_update_queued);
-                        queue_delayed_work(rds_wq, &conn->c_send_w, 0);
+                        rds_send_xmit(conn);
                }
        }
diff --git a/net/rds/connection.c b/net/rds/connection.c
index 7619b671ca28..870992e08cae 100644
--- a/net/rds/connection.c
+++ b/net/rds/connection.c
@@ -37,7 +37,6 @@
 #include "rds.h"
 #include "loop.h"
-#include "rdma.h"
 #define RDS_CONNECTION_HASH_BITS 12
 #define RDS_CONNECTION_HASH_ENTRIES (1 << RDS_CONNECTION_HASH_BITS)
@@ -63,18 +62,7 @@ static struct hlist_head *rds_conn_bucket(__be32 laddr, __be32 faddr)
                var |= RDS_INFO_CONNECTION_FLAG_##suffix;       \
 } while (0)
-static inline int rds_conn_is_sending(struct rds_connection *conn)
+/* rcu read lock must be held or the connection spinlock */
-{
-        int ret = 0;
-        if (!mutex_trylock(&conn->c_send_lock))
-                ret = 1;
-        else
-                mutex_unlock(&conn->c_send_lock);
-        return ret;
-}
 static struct rds_connection *rds_conn_lookup(struct hlist_head *head,
                                              __be32 laddr, __be32 faddr,
                                              struct rds_transport *trans)
@@ -82,7 +70,7 @@ static struct rds_connection *rds_conn_lookup(struct hlist_head *head,
        struct rds_connection *conn, *ret = NULL;
        struct hlist_node *pos;
-        hlist_for_each_entry(conn, pos, head, c_hash_node) {
+        hlist_for_each_entry_rcu(conn, pos, head, c_hash_node) {
                if (conn->c_faddr == faddr && conn->c_laddr == laddr &&
                                conn->c_trans == trans) {
                        ret = conn;
@@ -129,10 +117,11 @@ static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
 {
        struct rds_connection *conn, *parent = NULL;
        struct hlist_head *head = rds_conn_bucket(laddr, faddr);
+        struct rds_transport *loop_trans;
        unsigned long flags;
        int ret;
-        spin_lock_irqsave(&rds_conn_lock, flags);
+        rcu_read_lock();
        conn = rds_conn_lookup(head, laddr, faddr, trans);
        if (conn && conn->c_loopback && conn->c_trans != &rds_loop_transport &&
            !is_outgoing) {
@@ -143,12 +132,12 @@ static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
                parent = conn;
                conn = parent->c_passive;
        }
-        spin_unlock_irqrestore(&rds_conn_lock, flags);
+        rcu_read_unlock();
        if (conn)
                goto out;
        conn = kmem_cache_zalloc(rds_conn_slab, gfp);
-        if (conn == NULL) {
+        if (!conn) {
                conn = ERR_PTR(-ENOMEM);
                goto out;
        }
@@ -159,7 +148,7 @@ static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
        spin_lock_init(&conn->c_lock);
        conn->c_next_tx_seq = 1;
-        mutex_init(&conn->c_send_lock);
+        init_waitqueue_head(&conn->c_waitq);
        INIT_LIST_HEAD(&conn->c_send_queue);
        INIT_LIST_HEAD(&conn->c_retrans);
@@ -175,7 +164,9 @@ static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
         * can bind to the destination address then we'd rather the messages
         * flow through loopback rather than either transport.
         */
-        if (rds_trans_get_preferred(faddr)) {
+        loop_trans = rds_trans_get_preferred(faddr);
+        if (loop_trans) {
+                rds_trans_put(loop_trans);
                conn->c_loopback = 1;
                if (is_outgoing && trans->t_prefer_loopback) {
                        /* "outgoing" connection - and the transport
@@ -238,7 +229,7 @@ static struct rds_connection *__rds_conn_create(__be32 laddr, __be32 faddr,
                        kmem_cache_free(rds_conn_slab, conn);
                        conn = found;
                } else {
-                        hlist_add_head(&conn->c_hash_node, head);
+                        hlist_add_head_rcu(&conn->c_hash_node, head);
                        rds_cong_add_conn(conn);
                        rds_conn_count++;
                }
@@ -263,21 +254,91 @@ struct rds_connection *rds_conn_create_outgoing(__be32 laddr, __be32 faddr,
 }
 EXPORT_SYMBOL_GPL(rds_conn_create_outgoing);
+void rds_conn_shutdown(struct rds_connection *conn)
+{
+        /* shut it down unless it's down already */
+        if (!rds_conn_transition(conn, RDS_CONN_DOWN, RDS_CONN_DOWN)) {
+                /*
+                 * Quiesce the connection mgmt handlers before we start tearing
+                 * things down. We don't hold the mutex for the entire
+                 * duration of the shutdown operation, else we may be
+                 * deadlocking with the CM handler. Instead, the CM event
+                 * handler is supposed to check for state DISCONNECTING
+                 */
+                mutex_lock(&conn->c_cm_lock);
+                if (!rds_conn_transition(conn, RDS_CONN_UP, RDS_CONN_DISCONNECTING)
+                 && !rds_conn_transition(conn, RDS_CONN_ERROR, RDS_CONN_DISCONNECTING)) {
+                        rds_conn_error(conn, "shutdown called in state %d\n",
+                                        atomic_read(&conn->c_state));
+                        mutex_unlock(&conn->c_cm_lock);
+                        return;
+                }
+                mutex_unlock(&conn->c_cm_lock);
+                wait_event(conn->c_waitq,
+                           !test_bit(RDS_IN_XMIT, &conn->c_flags));
+                conn->c_trans->conn_shutdown(conn);
+                rds_conn_reset(conn);
+                if (!rds_conn_transition(conn, RDS_CONN_DISCONNECTING, RDS_CONN_DOWN)) {
+                        /* This can happen - eg when we're in the middle of tearing
+                         * down the connection, and someone unloads the rds module.
+                         * Quite reproduceable with loopback connections.
+                         * Mostly harmless.
+                         */
+                        rds_conn_error(conn,
+                                "%s: failed to transition to state DOWN, "
+                                "current state is %d\n",
+                                __func__,
+                                atomic_read(&conn->c_state));
+                        return;
+                }
+        }
+        /* Then reconnect if it's still live.
+         * The passive side of an IB loopback connection is never added
+         * to the conn hash, so we never trigger a reconnect on this
+         * conn - the reconnect is always triggered by the active peer. */
+        cancel_delayed_work_sync(&conn->c_conn_w);
+        rcu_read_lock();
+        if (!hlist_unhashed(&conn->c_hash_node)) {
+                rcu_read_unlock();
+                rds_queue_reconnect(conn);
+        } else {
+                rcu_read_unlock();
+        }
+}
+/*
+ * Stop and free a connection.
+ *
+ * This can only be used in very limited circumstances.  It assumes that once
+ * the conn has been shutdown that no one else is referencing the connection.
+ * We can only ensure this in the rmmod path in the current code.
+ */
 void rds_conn_destroy(struct rds_connection *conn)
 {
        struct rds_message *rm, *rtmp;
+        unsigned long flags;
        rdsdebug("freeing conn %p for %pI4 -> "
                 "%pI4\n", conn, &conn->c_laddr,
                 &conn->c_faddr);
-        hlist_del_init(&conn->c_hash_node);
+        /* Ensure conn will not be scheduled for reconnect */
+        spin_lock_irq(&rds_conn_lock);
+        hlist_del_init_rcu(&conn->c_hash_node);
+        spin_unlock_irq(&rds_conn_lock);
+        synchronize_rcu();
-        /* wait for the rds thread to shut it down */
+        /* shut the connection down */
-        atomic_set(&conn->c_state, RDS_CONN_ERROR);
+        rds_conn_drop(conn);
-        cancel_delayed_work(&conn->c_conn_w);
+        flush_work(&conn->c_down_w);
-        queue_work(rds_wq, &conn->c_down_w);
-        flush_workqueue(rds_wq);
+        /* make sure lingering queued work won't try to ref the conn */
+        cancel_delayed_work_sync(&conn->c_send_w);
+        cancel_delayed_work_sync(&conn->c_recv_w);
        /* tear down queued messages */
        list_for_each_entry_safe(rm, rtmp,
@@ -302,7 +363,9 @@ void rds_conn_destroy(struct rds_connection *conn)
        BUG_ON(!list_empty(&conn->c_retrans));
        kmem_cache_free(rds_conn_slab, conn);
+        spin_lock_irqsave(&rds_conn_lock, flags);
        rds_conn_count--;
+        spin_unlock_irqrestore(&rds_conn_lock, flags);
 }
 EXPORT_SYMBOL_GPL(rds_conn_destroy);
@@ -316,23 +379,23 @@ static void rds_conn_message_info(struct socket *sock, unsigned int len,
        struct list_head *list;
        struct rds_connection *conn;
        struct rds_message *rm;
-        unsigned long flags;
        unsigned int total = 0;
+        unsigned long flags;
        size_t i;
        len /= sizeof(struct rds_info_message);
-        spin_lock_irqsave(&rds_conn_lock, flags);
+        rcu_read_lock();
        for (i = 0, head = rds_conn_hash; i < ARRAY_SIZE(rds_conn_hash);
             i++, head++) {
-                hlist_for_each_entry(conn, pos, head, c_hash_node) {
+                hlist_for_each_entry_rcu(conn, pos, head, c_hash_node) {
                        if (want_send)
                                list = &conn->c_send_queue;
                        else
                                list = &conn->c_retrans;
-                        spin_lock(&conn->c_lock);
+                        spin_lock_irqsave(&conn->c_lock, flags);
                        /* XXX too lazy to maintain counts.. */
                        list_for_each_entry(rm, list, m_conn_item) {
@@ -343,11 +406,10 @@ static void rds_conn_message_info(struct socket *sock, unsigned int len,
                                                          conn->c_faddr, 0);
                        }
-                        spin_unlock(&conn->c_lock);
+                        spin_unlock_irqrestore(&conn->c_lock, flags);
                }
        }
+        rcu_read_unlock();
-        spin_unlock_irqrestore(&rds_conn_lock, flags);
        lens->nr = total;
        lens->each = sizeof(struct rds_info_message);
@@ -377,19 +439,17 @@ void rds_for_each_conn_info(struct socket *sock, unsigned int len,
        uint64_t buffer[(item_len + 7) / 8];
        struct hlist_head *head;
        struct hlist_node *pos;
-        struct hlist_node *tmp;
        struct rds_connection *conn;
-        unsigned long flags;
        size_t i;
-        spin_lock_irqsave(&rds_conn_lock, flags);
+        rcu_read_lock();
        lens->nr = 0;
        lens->each = item_len;
        for (i = 0, head = rds_conn_hash; i < ARRAY_SIZE(rds_conn_hash);
             i++, head++) {
-                hlist_for_each_entry_safe(conn, pos, tmp, head, c_hash_node) {
+                hlist_for_each_entry_rcu(conn, pos, head, c_hash_node) {
                        /* XXX no c_lock usage.. */
                        if (!visitor(conn, buffer))
@@ -405,8 +465,7 @@ void rds_for_each_conn_info(struct socket *sock, unsigned int len,
                        lens->nr++;
                }
        }
+        rcu_read_unlock();
-        spin_unlock_irqrestore(&rds_conn_lock, flags);
 }
 EXPORT_SYMBOL_GPL(rds_for_each_conn_info);
@@ -423,8 +482,8 @@ static int rds_conn_info_visitor(struct rds_connection *conn,
                sizeof(cinfo->transport));
        cinfo->flags = 0;
-        rds_conn_info_set(cinfo->flags,
+        rds_conn_info_set(cinfo->flags, test_bit(RDS_IN_XMIT, &conn->c_flags),
-                          rds_conn_is_sending(conn), SENDING);
+                          SENDING);
        /* XXX Future: return the state rather than these funky bits */
        rds_conn_info_set(cinfo->flags,
                          atomic_read(&conn->c_state) == RDS_CONN_CONNECTING,
@@ -444,12 +503,12 @@ static void rds_conn_info(struct socket *sock, unsigned int len,
                                sizeof(struct rds_info_connection));
 }
-int __init rds_conn_init(void)
+int rds_conn_init(void)
 {
        rds_conn_slab = kmem_cache_create("rds_connection",
                                          sizeof(struct rds_connection),
                                          0, 0, NULL);
-        if (rds_conn_slab == NULL)
+        if (!rds_conn_slab)
                return -ENOMEM;
        rds_info_register_func(RDS_INFO_CONNECTIONS, rds_conn_info);
@@ -487,6 +546,18 @@ void rds_conn_drop(struct rds_connection *conn)
 EXPORT_SYMBOL_GPL(rds_conn_drop);
 /*
+ * If the connection is down, trigger a connect. We may have scheduled a
+ * delayed reconnect however - in this case we should not interfere.
+ */
+void rds_conn_connect_if_down(struct rds_connection *conn)
+{
+        if (rds_conn_state(conn) == RDS_CONN_DOWN &&
+            !test_and_set_bit(RDS_RECONNECT_PENDING, &conn->c_flags))
+                queue_delayed_work(rds_wq, &conn->c_conn_w, 0);
+}
+EXPORT_SYMBOL_GPL(rds_conn_connect_if_down);
+/*
 * An error occurred on the connection
 */
 void
diff --git a/net/rds/ib.c b/net/rds/ib.c
index 8f2d6dd7700a..b12a3951167d 100644
--- a/net/rds/ib.c
+++ b/net/rds/ib.c
@@ -53,12 +53,71 @@ MODULE_PARM_DESC(fmr_message_size, " Max size of a RDMA transfer");
 module_param(rds_ib_retry_count, int, 0444);
 MODULE_PARM_DESC(rds_ib_retry_count, " Number of hw retries before reporting an error");
+/*
+ * we have a clumsy combination of RCU and a rwsem protecting this list
+ * because it is used both in the get_mr fast path and while blocking in
+ * the FMR flushing path.
+ */
+DECLARE_RWSEM(rds_ib_devices_lock);
 struct list_head rds_ib_devices;
 /* NOTE: if also grabbing ibdev lock, grab this first */
 DEFINE_SPINLOCK(ib_nodev_conns_lock);
 LIST_HEAD(ib_nodev_conns);
+void rds_ib_nodev_connect(void)
+{
+        struct rds_ib_connection *ic;
+        spin_lock(&ib_nodev_conns_lock);
+        list_for_each_entry(ic, &ib_nodev_conns, ib_node)
+                rds_conn_connect_if_down(ic->conn);
+        spin_unlock(&ib_nodev_conns_lock);
+}
+void rds_ib_dev_shutdown(struct rds_ib_device *rds_ibdev)
+{
+        struct rds_ib_connection *ic;
+        unsigned long flags;
+        spin_lock_irqsave(&rds_ibdev->spinlock, flags);
+        list_for_each_entry(ic, &rds_ibdev->conn_list, ib_node)
+                rds_conn_drop(ic->conn);
+        spin_unlock_irqrestore(&rds_ibdev->spinlock, flags);
+}
+/*
+ * rds_ib_destroy_mr_pool() blocks on a few things and mrs drop references
+ * from interrupt context so we push freing off into a work struct in krdsd.
+ */
+static void rds_ib_dev_free(struct work_struct *work)
+{
+        struct rds_ib_ipaddr *i_ipaddr, *i_next;
+        struct rds_ib_device *rds_ibdev = container_of(work,
+                                        struct rds_ib_device, free_work);
+        if (rds_ibdev->mr_pool)
+                rds_ib_destroy_mr_pool(rds_ibdev->mr_pool);
+        if (rds_ibdev->mr)
+                ib_dereg_mr(rds_ibdev->mr);
+        if (rds_ibdev->pd)
+                ib_dealloc_pd(rds_ibdev->pd);
+        list_for_each_entry_safe(i_ipaddr, i_next, &rds_ibdev->ipaddr_list, list) {
+                list_del(&i_ipaddr->list);
+                kfree(i_ipaddr);
+        }
+        kfree(rds_ibdev);
+}
+void rds_ib_dev_put(struct rds_ib_device *rds_ibdev)
+{
+        BUG_ON(atomic_read(&rds_ibdev->refcount) <= 0);
+        if (atomic_dec_and_test(&rds_ibdev->refcount))
+                queue_work(rds_wq, &rds_ibdev->free_work);
+}
 void rds_ib_add_one(struct ib_device *device)
 {
        struct rds_ib_device *rds_ibdev;
@@ -77,11 +136,14 @@ void rds_ib_add_one(struct ib_device *device)
                goto free_attr;
        }
-        rds_ibdev = kmalloc(sizeof *rds_ibdev, GFP_KERNEL);
+        rds_ibdev = kzalloc_node(sizeof(struct rds_ib_device), GFP_KERNEL,
+                                 ibdev_to_node(device));
        if (!rds_ibdev)
                goto free_attr;
        spin_lock_init(&rds_ibdev->spinlock);
+        atomic_set(&rds_ibdev->refcount, 1);
+        INIT_WORK(&rds_ibdev->free_work, rds_ib_dev_free);
        rds_ibdev->max_wrs = dev_attr->max_qp_wr;
        rds_ibdev->max_sge = min(dev_attr->max_sge, RDS_IB_MAX_SGE);
@@ -91,68 +153,107 @@ void rds_ib_add_one(struct ib_device *device)
                        min_t(unsigned int, dev_attr->max_fmr, fmr_pool_size) :
                        fmr_pool_size;
+        rds_ibdev->max_initiator_depth = dev_attr->max_qp_init_rd_atom;
+        rds_ibdev->max_responder_resources = dev_attr->max_qp_rd_atom;
        rds_ibdev->dev = device;
        rds_ibdev->pd = ib_alloc_pd(device);
-        if (IS_ERR(rds_ibdev->pd))
+        if (IS_ERR(rds_ibdev->pd)) {
-                goto free_dev;
+                rds_ibdev->pd = NULL;
+                goto put_dev;
+        }
-        rds_ibdev->mr = ib_get_dma_mr(rds_ibdev->pd,
+        rds_ibdev->mr = ib_get_dma_mr(rds_ibdev->pd, IB_ACCESS_LOCAL_WRITE);
-                                      IB_ACCESS_LOCAL_WRITE);
+        if (IS_ERR(rds_ibdev->mr)) {
-        if (IS_ERR(rds_ibdev->mr))
+                rds_ibdev->mr = NULL;
-                goto err_pd;
+                goto put_dev;
+        }
        rds_ibdev->mr_pool = rds_ib_create_mr_pool(rds_ibdev);
        if (IS_ERR(rds_ibdev->mr_pool)) {
                rds_ibdev->mr_pool = NULL;
-                goto err_mr;
+                goto put_dev;
        }
        INIT_LIST_HEAD(&rds_ibdev->ipaddr_list);
        INIT_LIST_HEAD(&rds_ibdev->conn_list);
-        list_add_tail(&rds_ibdev->list, &rds_ib_devices);
+        down_write(&rds_ib_devices_lock);
+        list_add_tail_rcu(&rds_ibdev->list, &rds_ib_devices);
+        up_write(&rds_ib_devices_lock);
+        atomic_inc(&rds_ibdev->refcount);
        ib_set_client_data(device, &rds_ib_client, rds_ibdev);
+        atomic_inc(&rds_ibdev->refcount);
-        goto free_attr;
+        rds_ib_nodev_connect();
-err_mr:
+put_dev:
-        ib_dereg_mr(rds_ibdev->mr);
+        rds_ib_dev_put(rds_ibdev);
-err_pd:
-        ib_dealloc_pd(rds_ibdev->pd);
-free_dev:
-        kfree(rds_ibdev);
 free_attr:
        kfree(dev_attr);
 }
+/*
+ * New connections use this to find the device to associate with the
+ * connection.  It's not in the fast path so we're not concerned about the
+ * performance of the IB call.  (As of this writing, it uses an interrupt
+ * blocking spinlock to serialize walking a per-device list of all registered
+ * clients.)
+ *
+ * RCU is used to handle incoming connections racing with device teardown.
+ * Rather than use a lock to serialize removal from the client_data and
+ * getting a new reference, we use an RCU grace period.  The destruction
+ * path removes the device from client_data and then waits for all RCU
+ * readers to finish.
+ *
+ * A new connection can get NULL from this if its arriving on a
+ * device that is in the process of being removed.
+ */
+struct rds_ib_device *rds_ib_get_client_data(struct ib_device *device)
+{
+        struct rds_ib_device *rds_ibdev;
+        rcu_read_lock();
+        rds_ibdev = ib_get_client_data(device, &rds_ib_client);
+        if (rds_ibdev)
+                atomic_inc(&rds_ibdev->refcount);
+        rcu_read_unlock();
+        return rds_ibdev;
+}
+/*
+ * The IB stack is letting us know that a device is going away.  This can
+ * happen if the underlying HCA driver is removed or if PCI hotplug is removing
+ * the pci function, for example.
+ *
+ * This can be called at any time and can be racing with any other RDS path.
+ */
 void rds_ib_remove_one(struct ib_device *device)
 {
        struct rds_ib_device *rds_ibdev;
-        struct rds_ib_ipaddr *i_ipaddr, *i_next;
        rds_ibdev = ib_get_client_data(device, &rds_ib_client);
        if (!rds_ibdev)
                return;
-        list_for_each_entry_safe(i_ipaddr, i_next, &rds_ibdev->ipaddr_list, list) {
+        rds_ib_dev_shutdown(rds_ibdev);
-                list_del(&i_ipaddr->list);
-                kfree(i_ipaddr);
-        }
-        rds_ib_destroy_conns(rds_ibdev);
+        /* stop connection attempts from getting a reference to this device. */
+        ib_set_client_data(device, &rds_ib_client, NULL);
-        if (rds_ibdev->mr_pool)
+        down_write(&rds_ib_devices_lock);
-                rds_ib_destroy_mr_pool(rds_ibdev->mr_pool);
+        list_del_rcu(&rds_ibdev->list);
+        up_write(&rds_ib_devices_lock);
-        ib_dereg_mr(rds_ibdev->mr);
-        while (ib_dealloc_pd(rds_ibdev->pd)) {
-                rdsdebug("Failed to dealloc pd %p\n", rds_ibdev->pd);
-                msleep(1);
-        }
-        list_del(&rds_ibdev->list);
+        /*
-        kfree(rds_ibdev);
+         * This synchronize rcu is waiting for readers of both the ib
+         * client data and the devices list to finish before we drop
+         * both of those references.
+         */
+        synchronize_rcu();
+        rds_ib_dev_put(rds_ibdev);
+        rds_ib_dev_put(rds_ibdev);
 }
 struct ib_client rds_ib_client = {
@@ -186,7 +287,7 @@ static int rds_ib_conn_info_visitor(struct rds_connection *conn,
                rdma_addr_get_sgid(dev_addr, (union ib_gid *) &iinfo->src_gid);
                rdma_addr_get_dgid(dev_addr, (union ib_gid *) &iinfo->dst_gid);
-                rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
+                rds_ibdev = ic->rds_ibdev;
                iinfo->max_send_wr = ic->i_send_ring.w_nr;
                iinfo->max_recv_wr = ic->i_recv_ring.w_nr;
                iinfo->max_send_sge = rds_ibdev->max_sge;
@@ -248,29 +349,36 @@ static int rds_ib_laddr_check(__be32 addr)
        return ret;
 }
+static void rds_ib_unregister_client(void)
+{
+        ib_unregister_client(&rds_ib_client);
+        /* wait for rds_ib_dev_free() to complete */
+        flush_workqueue(rds_wq);
+}
 void rds_ib_exit(void)
 {
        rds_info_deregister_func(RDS_INFO_IB_CONNECTIONS, rds_ib_ic_info);
+        rds_ib_unregister_client();
        rds_ib_destroy_nodev_conns();
-        ib_unregister_client(&rds_ib_client);
        rds_ib_sysctl_exit();
        rds_ib_recv_exit();
        rds_trans_unregister(&rds_ib_transport);
+        rds_ib_fmr_exit();
 }
 struct rds_transport rds_ib_transport = {
        .laddr_check            = rds_ib_laddr_check,
        .xmit_complete          = rds_ib_xmit_complete,
        .xmit                   = rds_ib_xmit,
-        .xmit_cong_map          = NULL,
        .xmit_rdma              = rds_ib_xmit_rdma,
+        .xmit_atomic            = rds_ib_xmit_atomic,
        .recv                   = rds_ib_recv,
        .conn_alloc             = rds_ib_conn_alloc,
        .conn_free              = rds_ib_conn_free,
        .conn_connect           = rds_ib_conn_connect,
        .conn_shutdown          = rds_ib_conn_shutdown,
        .inc_copy_to_user       = rds_ib_inc_copy_to_user,
-        .inc_purge              = rds_ib_inc_purge,
        .inc_free               = rds_ib_inc_free,
        .cm_initiate_connect    = rds_ib_cm_initiate_connect,
        .cm_handle_connect      = rds_ib_cm_handle_connect,
@@ -286,16 +394,20 @@ struct rds_transport rds_ib_transport = {
        .t_type                 = RDS_TRANS_IB
 };
-int __init rds_ib_init(void)
+int rds_ib_init(void)
 {
        int ret;
        INIT_LIST_HEAD(&rds_ib_devices);
-        ret = ib_register_client(&rds_ib_client);
+        ret = rds_ib_fmr_init();
        if (ret)
                goto out;
+        ret = ib_register_client(&rds_ib_client);
+        if (ret)
+                goto out_fmr_exit;
        ret = rds_ib_sysctl_init();
        if (ret)
                goto out_ibreg;
@@ -317,7 +429,9 @@ out_recv:
 out_sysctl:
        rds_ib_sysctl_exit();
 out_ibreg:
-        ib_unregister_client(&rds_ib_client);
+        rds_ib_unregister_client();
+out_fmr_exit:
+        rds_ib_fmr_exit();
 out:
        return ret;
 }
diff --git a/net/rds/ib.h b/net/rds/ib.h
index 64df4e79b29f..7ad3d57e06a5 100644
--- a/net/rds/ib.h
+++ b/net/rds/ib.h
@@ -3,11 +3,13 @@
 #include <rdma/ib_verbs.h>
 #include <rdma/rdma_cm.h>
+#include <linux/pci.h>
+#include <linux/slab.h>
 #include "rds.h"
 #include "rdma_transport.h"
 #define RDS_FMR_SIZE                    256
-#define RDS_FMR_POOL_SIZE               4096
+#define RDS_FMR_POOL_SIZE               8192
 #define RDS_IB_MAX_SGE                  8
 #define RDS_IB_RECV_SGE                 2
@@ -19,6 +21,9 @@
 #define RDS_IB_SUPPORTED_PROTOCOLS      0x00000003      /* minor versions supported */
+#define RDS_IB_RECYCLE_BATCH_COUNT      32
+extern struct rw_semaphore rds_ib_devices_lock;
 extern struct list_head rds_ib_devices;
 /*
@@ -26,20 +31,29 @@ extern struct list_head rds_ib_devices;
 * try and minimize the amount of memory tied up both the device and
 * socket receive queues.
 */
-/* page offset of the final full frag that fits in the page */
-#define RDS_PAGE_LAST_OFF (((PAGE_SIZE  / RDS_FRAG_SIZE) - 1) * RDS_FRAG_SIZE)
 struct rds_page_frag {
        struct list_head        f_item;
-        struct page             *f_page;
+        struct list_head        f_cache_entry;
-        unsigned long           f_offset;
+        struct scatterlist      f_sg;
-        dma_addr_t              f_mapped;
 };
 struct rds_ib_incoming {
        struct list_head        ii_frags;
+        struct list_head        ii_cache_entry;
        struct rds_incoming     ii_inc;
 };
+struct rds_ib_cache_head {
+        struct list_head *first;
+        unsigned long count;
+};
+struct rds_ib_refill_cache {
+        struct rds_ib_cache_head *percpu;
+        struct list_head         *xfer;
+        struct list_head         *ready;
+};
 struct rds_ib_connect_private {
        /* Add new fields at the end, and don't permute existing fields. */
        __be32                  dp_saddr;
@@ -53,8 +67,7 @@ struct rds_ib_connect_private {
 };
 struct rds_ib_send_work {
-        struct rds_message      *s_rm;
+        void                    *s_op;
-        struct rds_rdma_op      *s_op;
        struct ib_send_wr       s_wr;
        struct ib_sge           s_sge[RDS_IB_MAX_SGE];
        unsigned long           s_queued;
@@ -92,10 +105,11 @@ struct rds_ib_connection {
        /* tx */
        struct rds_ib_work_ring i_send_ring;
-        struct rds_message      *i_rm;
+        struct rm_data_op       *i_data_op;
        struct rds_header       *i_send_hdrs;
        u64                     i_send_hdrs_dma;
        struct rds_ib_send_work *i_sends;
+        atomic_t                i_signaled_sends;
        /* rx */
        struct tasklet_struct   i_recv_tasklet;
@@ -106,8 +120,9 @@ struct rds_ib_connection {
        struct rds_header       *i_recv_hdrs;
        u64                     i_recv_hdrs_dma;
        struct rds_ib_recv_work *i_recvs;
-        struct rds_page_frag    i_frag;
        u64                     i_ack_recv;     /* last ACK received */
+        struct rds_ib_refill_cache i_cache_incs;
+        struct rds_ib_refill_cache i_cache_frags;
        /* sending acks */
        unsigned long           i_ack_flags;
@@ -138,7 +153,6 @@ struct rds_ib_connection {
        /* Batched completions */
        unsigned int            i_unsignaled_wrs;
-        long                    i_unsignaled_bytes;
 };
 /* This assumes that atomic_t is at least 32 bits */
@@ -164,9 +178,17 @@ struct rds_ib_device {
        unsigned int            max_fmrs;
        int                     max_sge;
        unsigned int            max_wrs;
+        unsigned int            max_initiator_depth;
+        unsigned int            max_responder_resources;
        spinlock_t              spinlock;       /* protect the above */
+        atomic_t                refcount;
+        struct work_struct      free_work;
 };
+#define pcidev_to_node(pcidev) pcibus_to_node(pcidev->bus)
+#define ibdev_to_node(ibdev) pcidev_to_node(to_pci_dev(ibdev->dma_device))
+#define rdsibdev_to_node(rdsibdev) ibdev_to_node(rdsibdev->dev)
 /* bits for i_ack_flags */
 #define IB_ACK_IN_FLIGHT        0
 #define IB_ACK_REQUESTED        1
@@ -202,6 +224,8 @@ struct rds_ib_statistics {
        uint64_t        s_ib_rdma_mr_pool_flush;
        uint64_t        s_ib_rdma_mr_pool_wait;
        uint64_t        s_ib_rdma_mr_pool_depleted;
+        uint64_t        s_ib_atomic_cswp;
+        uint64_t        s_ib_atomic_fadd;
 };
 extern struct workqueue_struct *rds_ib_wq;
@@ -243,6 +267,8 @@ static inline void rds_ib_dma_sync_sg_for_device(struct ib_device *dev,
 extern struct rds_transport rds_ib_transport;
 extern void rds_ib_add_one(struct ib_device *device);
 extern void rds_ib_remove_one(struct ib_device *device);
+struct rds_ib_device *rds_ib_get_client_data(struct ib_device *device);
+void rds_ib_dev_put(struct rds_ib_device *rds_ibdev);
 extern struct ib_client rds_ib_client;
 extern unsigned int fmr_pool_size;
@@ -258,7 +284,7 @@ void rds_ib_conn_free(void *arg);
 int rds_ib_conn_connect(struct rds_connection *conn);
 void rds_ib_conn_shutdown(struct rds_connection *conn);
 void rds_ib_state_change(struct sock *sk);
-int __init rds_ib_listen_init(void);
+int rds_ib_listen_init(void);
 void rds_ib_listen_stop(void);
 void __rds_ib_conn_error(struct rds_connection *conn, const char *, ...);
 int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
@@ -275,15 +301,7 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn,
 int rds_ib_update_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr);
 void rds_ib_add_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
 void rds_ib_remove_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn);
-void __rds_ib_destroy_conns(struct list_head *list, spinlock_t *list_lock);
+void rds_ib_destroy_nodev_conns(void);
-static inline void rds_ib_destroy_nodev_conns(void)
-{
-        __rds_ib_destroy_conns(&ib_nodev_conns, &ib_nodev_conns_lock);
-}
-static inline void rds_ib_destroy_conns(struct rds_ib_device *rds_ibdev)
-{
-        __rds_ib_destroy_conns(&rds_ibdev->conn_list, &rds_ibdev->spinlock);
-}
 struct rds_ib_mr_pool *rds_ib_create_mr_pool(struct rds_ib_device *);
 void rds_ib_get_mr_info(struct rds_ib_device *rds_ibdev, struct rds_info_rdma_connection *iinfo);
 void rds_ib_destroy_mr_pool(struct rds_ib_mr_pool *);
@@ -292,14 +310,16 @@ void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
 void rds_ib_sync_mr(void *trans_private, int dir);
 void rds_ib_free_mr(void *trans_private, int invalidate);
 void rds_ib_flush_mrs(void);
+int rds_ib_fmr_init(void);
+void rds_ib_fmr_exit(void);
 /* ib_recv.c */
-int __init rds_ib_recv_init(void);
+int rds_ib_recv_init(void);
 void rds_ib_recv_exit(void);
 int rds_ib_recv(struct rds_connection *conn);
-int rds_ib_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
+int rds_ib_recv_alloc_caches(struct rds_ib_connection *ic);
-                       gfp_t page_gfp, int prefill);
+void rds_ib_recv_free_caches(struct rds_ib_connection *ic);
-void rds_ib_inc_purge(struct rds_incoming *inc);
+void rds_ib_recv_refill(struct rds_connection *conn, int prefill);
 void rds_ib_inc_free(struct rds_incoming *inc);
 int rds_ib_inc_copy_to_user(struct rds_incoming *inc, struct iovec *iov,
                             size_t size);
@@ -325,17 +345,19 @@ u32 rds_ib_ring_completed(struct rds_ib_work_ring *ring, u32 wr_id, u32 oldest);
 extern wait_queue_head_t rds_ib_ring_empty_wait;
 /* ib_send.c */
+char *rds_ib_wc_status_str(enum ib_wc_status status);
 void rds_ib_xmit_complete(struct rds_connection *conn);
 int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
                unsigned int hdr_off, unsigned int sg, unsigned int off);
 void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context);
 void rds_ib_send_init_ring(struct rds_ib_connection *ic);
 void rds_ib_send_clear_ring(struct rds_ib_connection *ic);
-int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op);
+int rds_ib_xmit_rdma(struct rds_connection *conn, struct rm_rdma_op *op);
 void rds_ib_send_add_credits(struct rds_connection *conn, unsigned int credits);
 void rds_ib_advertise_credits(struct rds_connection *conn, unsigned int posted);
 int rds_ib_send_grab_credits(struct rds_ib_connection *ic, u32 wanted,
                             u32 *adv_credits, int need_posted, int max_posted);
+int rds_ib_xmit_atomic(struct rds_connection *conn, struct rm_atomic_op *op);
 /* ib_stats.c */
 DECLARE_PER_CPU(struct rds_ib_statistics, rds_ib_stats);
@@ -344,7 +366,7 @@ unsigned int rds_ib_stats_info_copy(struct rds_info_iterator *iter,
                                    unsigned int avail);
 /* ib_sysctl.c */
-int __init rds_ib_sysctl_init(void);
+int rds_ib_sysctl_init(void);
 void rds_ib_sysctl_exit(void);
 extern unsigned long rds_ib_sysctl_max_send_wr;
 extern unsigned long rds_ib_sysctl_max_recv_wr;
@@ -354,28 +376,4 @@ extern unsigned long rds_ib_sysctl_max_recv_allocation;
 extern unsigned int rds_ib_sysctl_flow_control;
 extern ctl_table rds_ib_sysctl_table[];
-/*
- * Helper functions for getting/setting the header and data SGEs in
- * RDS packets (not RDMA)
- *
- * From version 3.1 onwards, header is in front of data in the sge.
- */
-static inline struct ib_sge *
-rds_ib_header_sge(struct rds_ib_connection *ic, struct ib_sge *sge)
-{
-        if (ic->conn->c_version > RDS_PROTOCOL_3_0)
-                return &sge[0];
-        else
-                return &sge[1];
-}
-static inline struct ib_sge *
-rds_ib_data_sge(struct rds_ib_connection *ic, struct ib_sge *sge)
-{
-        if (ic->conn->c_version > RDS_PROTOCOL_3_0)
-                return &sge[1];
-        else
-                return &sge[0];
-}
 #endif
diff --git a/net/rds/ib_cm.c b/net/rds/ib_cm.c
index f68832798db2..ee369d201a65 100644
--- a/net/rds/ib_cm.c
+++ b/net/rds/ib_cm.c
@@ -38,6 +38,36 @@
 #include "rds.h"
 #include "ib.h"
+static char *rds_ib_event_type_strings[] = {
+#define RDS_IB_EVENT_STRING(foo) \
+                [IB_EVENT_##foo] = __stringify(IB_EVENT_##foo)
+        RDS_IB_EVENT_STRING(CQ_ERR),
+        RDS_IB_EVENT_STRING(QP_FATAL),
+        RDS_IB_EVENT_STRING(QP_REQ_ERR),
+        RDS_IB_EVENT_STRING(QP_ACCESS_ERR),
+        RDS_IB_EVENT_STRING(COMM_EST),
+        RDS_IB_EVENT_STRING(SQ_DRAINED),
+        RDS_IB_EVENT_STRING(PATH_MIG),
+        RDS_IB_EVENT_STRING(PATH_MIG_ERR),
+        RDS_IB_EVENT_STRING(DEVICE_FATAL),
+        RDS_IB_EVENT_STRING(PORT_ACTIVE),
+        RDS_IB_EVENT_STRING(PORT_ERR),
+        RDS_IB_EVENT_STRING(LID_CHANGE),
+        RDS_IB_EVENT_STRING(PKEY_CHANGE),
+        RDS_IB_EVENT_STRING(SM_CHANGE),
+        RDS_IB_EVENT_STRING(SRQ_ERR),
+        RDS_IB_EVENT_STRING(SRQ_LIMIT_REACHED),
+        RDS_IB_EVENT_STRING(QP_LAST_WQE_REACHED),
+        RDS_IB_EVENT_STRING(CLIENT_REREGISTER),
+#undef RDS_IB_EVENT_STRING
+};
+static char *rds_ib_event_str(enum ib_event_type type)
+{
+        return rds_str_array(rds_ib_event_type_strings,
+                             ARRAY_SIZE(rds_ib_event_type_strings), type);
+};
 /*
 * Set the selected protocol version
 */
@@ -95,7 +125,6 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
 {
        const struct rds_ib_connect_private *dp = NULL;
        struct rds_ib_connection *ic = conn->c_transport_data;
-        struct rds_ib_device *rds_ibdev;
        struct ib_qp_attr qp_attr;
        int err;
@@ -111,11 +140,21 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
                }
        }
-        printk(KERN_NOTICE "RDS/IB: connected to %pI4 version %u.%u%s\n",
+        if (conn->c_version < RDS_PROTOCOL(3,1)) {
-                        &conn->c_faddr,
+                printk(KERN_NOTICE "RDS/IB: Connection to %pI4 version %u.%u failed,"
-                        RDS_PROTOCOL_MAJOR(conn->c_version),
+                       " no longer supported\n",
-                        RDS_PROTOCOL_MINOR(conn->c_version),
+                       &conn->c_faddr,
-                        ic->i_flowctl ? ", flow control" : "");
+                       RDS_PROTOCOL_MAJOR(conn->c_version),
+                       RDS_PROTOCOL_MINOR(conn->c_version));
+                rds_conn_destroy(conn);
+                return;
+        } else {
+                printk(KERN_NOTICE "RDS/IB: connected to %pI4 version %u.%u%s\n",
+                       &conn->c_faddr,
+                       RDS_PROTOCOL_MAJOR(conn->c_version),
+                       RDS_PROTOCOL_MINOR(conn->c_version),
+                       ic->i_flowctl ? ", flow control" : "");
+        }
        /*
         * Init rings and fill recv. this needs to wait until protocol negotiation
@@ -125,7 +164,7 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
        rds_ib_recv_init_ring(ic);
        /* Post receive buffers - as a side effect, this will update
         * the posted credit count. */
-        rds_ib_recv_refill(conn, GFP_KERNEL, GFP_HIGHUSER, 1);
+        rds_ib_recv_refill(conn, 1);
        /* Tune RNR behavior */
        rds_ib_tune_rnr(ic, &qp_attr);
@@ -135,12 +174,11 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
        if (err)
                printk(KERN_NOTICE "ib_modify_qp(IB_QP_STATE, RTS): err=%d\n", err);
-        /* update ib_device with this local ipaddr & conn */
+        /* update ib_device with this local ipaddr */
-        rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
+        err = rds_ib_update_ipaddr(ic->rds_ibdev, conn->c_laddr);
-        err = rds_ib_update_ipaddr(rds_ibdev, conn->c_laddr);
        if (err)
-                printk(KERN_ERR "rds_ib_update_ipaddr failed (%d)\n", err);
+                printk(KERN_ERR "rds_ib_update_ipaddr failed (%d)\n",
-        rds_ib_add_conn(rds_ibdev, conn);
+                        err);
        /* If the peer gave us the last packet it saw, process this as if
         * we had received a regular ACK. */
@@ -153,18 +191,23 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
 static void rds_ib_cm_fill_conn_param(struct rds_connection *conn,
                        struct rdma_conn_param *conn_param,
                        struct rds_ib_connect_private *dp,
-                        u32 protocol_version)
+                        u32 protocol_version,
+                        u32 max_responder_resources,
+                        u32 max_initiator_depth)
 {
+        struct rds_ib_connection *ic = conn->c_transport_data;
+        struct rds_ib_device *rds_ibdev = ic->rds_ibdev;
        memset(conn_param, 0, sizeof(struct rdma_conn_param));
-        /* XXX tune these? */
-        conn_param->responder_resources = 1;
+        conn_param->responder_resources =
-        conn_param->initiator_depth = 1;
+                min_t(u32, rds_ibdev->max_responder_resources, max_responder_resources);
+        conn_param->initiator_depth =
+                min_t(u32, rds_ibdev->max_initiator_depth, max_initiator_depth);
        conn_param->retry_count = min_t(unsigned int, rds_ib_retry_count, 7);
        conn_param->rnr_retry_count = 7;
        if (dp) {
-                struct rds_ib_connection *ic = conn->c_transport_data;
                memset(dp, 0, sizeof(*dp));
                dp->dp_saddr = conn->c_laddr;
                dp->dp_daddr = conn->c_faddr;
@@ -189,7 +232,8 @@ static void rds_ib_cm_fill_conn_param(struct rds_connection *conn,
 static void rds_ib_cq_event_handler(struct ib_event *event, void *data)
 {
-        rdsdebug("event %u data %p\n", event->event, data);
+        rdsdebug("event %u (%s) data %p\n",
+                 event->event, rds_ib_event_str(event->event), data);
 }
 static void rds_ib_qp_event_handler(struct ib_event *event, void *data)
@@ -197,16 +241,18 @@ static void rds_ib_qp_event_handler(struct ib_event *event, void *data)
        struct rds_connection *conn = data;
        struct rds_ib_connection *ic = conn->c_transport_data;
-        rdsdebug("conn %p ic %p event %u\n", conn, ic, event->event);
+        rdsdebug("conn %p ic %p event %u (%s)\n", conn, ic, event->event,
+                 rds_ib_event_str(event->event));
        switch (event->event) {
        case IB_EVENT_COMM_EST:
                rdma_notify(ic->i_cm_id, IB_EVENT_COMM_EST);
                break;
        default:
-                rdsdebug("Fatal QP Event %u "
+                rdsdebug("Fatal QP Event %u (%s) "
                        "- connection %pI4->%pI4, reconnecting\n",
-                        event->event, &conn->c_laddr, &conn->c_faddr);
+                        event->event, rds_ib_event_str(event->event),
+                        &conn->c_laddr, &conn->c_faddr);
                rds_conn_drop(conn);
                break;
        }
@@ -224,18 +270,16 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
        struct rds_ib_device *rds_ibdev;
        int ret;
-        /* rds_ib_add_one creates a rds_ib_device object per IB device,
+        /*
-         * and allocates a protection domain, memory range and FMR pool
+         * It's normal to see a null device if an incoming connection races
-         * for each.  If that fails for any reason, it will not register
+         * with device removal, so we don't print a warning.
-         * the rds_ibdev at all.
         */
-        rds_ibdev = ib_get_client_data(dev, &rds_ib_client);
+        rds_ibdev = rds_ib_get_client_data(dev);
-        if (rds_ibdev == NULL) {
+        if (!rds_ibdev)
-                if (printk_ratelimit())
-                        printk(KERN_NOTICE "RDS/IB: No client_data for device %s\n",
-                                        dev->name);
                return -EOPNOTSUPP;
-        }
+        /* add the conn now so that connection establishment has the dev */
+        rds_ib_add_conn(rds_ibdev, conn);
        if (rds_ibdev->max_wrs < ic->i_send_ring.w_nr + 1)
                rds_ib_ring_resize(&ic->i_send_ring, rds_ibdev->max_wrs - 1);
@@ -306,7 +350,7 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
                                           ic->i_send_ring.w_nr *
                                                sizeof(struct rds_header),
                                           &ic->i_send_hdrs_dma, GFP_KERNEL);
-        if (ic->i_send_hdrs == NULL) {
+        if (!ic->i_send_hdrs) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent send failed\n");
                goto out;
@@ -316,7 +360,7 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
                                           ic->i_recv_ring.w_nr *
                                                sizeof(struct rds_header),
                                           &ic->i_recv_hdrs_dma, GFP_KERNEL);
-        if (ic->i_recv_hdrs == NULL) {
+        if (!ic->i_recv_hdrs) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent recv failed\n");
                goto out;
@@ -324,22 +368,24 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
        ic->i_ack = ib_dma_alloc_coherent(dev, sizeof(struct rds_header),
                                       &ic->i_ack_dma, GFP_KERNEL);
-        if (ic->i_ack == NULL) {
+        if (!ic->i_ack) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent ack failed\n");
                goto out;
        }
-        ic->i_sends = vmalloc(ic->i_send_ring.w_nr * sizeof(struct rds_ib_send_work));
+        ic->i_sends = vmalloc_node(ic->i_send_ring.w_nr * sizeof(struct rds_ib_send_work),
-        if (ic->i_sends == NULL) {
+                                   ibdev_to_node(dev));
+        if (!ic->i_sends) {
                ret = -ENOMEM;
                rdsdebug("send allocation failed\n");
                goto out;
        }
        memset(ic->i_sends, 0, ic->i_send_ring.w_nr * sizeof(struct rds_ib_send_work));
-        ic->i_recvs = vmalloc(ic->i_recv_ring.w_nr * sizeof(struct rds_ib_recv_work));
+        ic->i_recvs = vmalloc_node(ic->i_recv_ring.w_nr * sizeof(struct rds_ib_recv_work),
-        if (ic->i_recvs == NULL) {
+                                   ibdev_to_node(dev));
+        if (!ic->i_recvs) {
                ret = -ENOMEM;
                rdsdebug("recv allocation failed\n");
                goto out;
@@ -352,6 +398,7 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
                 ic->i_send_cq, ic->i_recv_cq);
 out:
+        rds_ib_dev_put(rds_ibdev);
        return ret;
 }
@@ -409,7 +456,7 @@ int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
        struct rds_ib_connection *ic = NULL;
        struct rdma_conn_param conn_param;
        u32 version;
-        int err, destroy = 1;
+        int err = 1, destroy = 1;
        /* Check whether the remote protocol version matches ours. */
        version = rds_ib_protocol_compatible(event);
@@ -448,7 +495,6 @@ int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
                        /* Wait and see - our connect may still be succeeding */
                        rds_ib_stats_inc(s_ib_connect_raced);
                }
-                mutex_unlock(&conn->c_cm_lock);
                goto out;
        }
@@ -475,24 +521,23 @@ int rds_ib_cm_handle_connect(struct rdma_cm_id *cm_id,
        err = rds_ib_setup_qp(conn);
        if (err) {
                rds_ib_conn_error(conn, "rds_ib_setup_qp failed (%d)\n", err);
-                mutex_unlock(&conn->c_cm_lock);
                goto out;
        }
-        rds_ib_cm_fill_conn_param(conn, &conn_param, &dp_rep, version);
+        rds_ib_cm_fill_conn_param(conn, &conn_param, &dp_rep, version,
+                event->param.conn.responder_resources,
+                event->param.conn.initiator_depth);
        /* rdma_accept() calls rdma_reject() internally if it fails */
        err = rdma_accept(cm_id, &conn_param);
-        mutex_unlock(&conn->c_cm_lock);
+        if (err)
-        if (err) {
                rds_ib_conn_error(conn, "rdma_accept failed (%d)\n", err);
-                goto out;
-        }
-        return 0;
 out:
-        rdma_reject(cm_id, NULL, 0);
+        if (conn)
+                mutex_unlock(&conn->c_cm_lock);
+        if (err)
+                rdma_reject(cm_id, NULL, 0);
        return destroy;
 }
@@ -516,8 +561,8 @@ int rds_ib_cm_initiate_connect(struct rdma_cm_id *cm_id)
                goto out;
        }
-        rds_ib_cm_fill_conn_param(conn, &conn_param, &dp, RDS_PROTOCOL_VERSION);
+        rds_ib_cm_fill_conn_param(conn, &conn_param, &dp, RDS_PROTOCOL_VERSION,
+                UINT_MAX, UINT_MAX);
        ret = rdma_connect(cm_id, &conn_param);
        if (ret)
                rds_ib_conn_error(conn, "rdma_connect failed (%d)\n", ret);
@@ -601,9 +646,19 @@ void rds_ib_conn_shutdown(struct rds_connection *conn)
                                ic->i_cm_id, err);
                }
+                /*
+                 * We want to wait for tx and rx completion to finish
+                 * before we tear down the connection, but we have to be
+                 * careful not to get stuck waiting on a send ring that
+                 * only has unsignaled sends in it.  We've shutdown new
+                 * sends before getting here so by waiting for signaled
+                 * sends to complete we're ensured that there will be no
+                 * more tx processing.
+                 */
                wait_event(rds_ib_ring_empty_wait,
-                        rds_ib_ring_empty(&ic->i_send_ring) &&
+                           rds_ib_ring_empty(&ic->i_recv_ring) &&
-                        rds_ib_ring_empty(&ic->i_recv_ring));
+                           (atomic_read(&ic->i_signaled_sends) == 0));
+                tasklet_kill(&ic->i_recv_tasklet);
                if (ic->i_send_hdrs)
                        ib_dma_free_coherent(dev,
@@ -654,9 +709,12 @@ void rds_ib_conn_shutdown(struct rds_connection *conn)
        BUG_ON(ic->rds_ibdev);
        /* Clear pending transmit */
-        if (ic->i_rm) {
+        if (ic->i_data_op) {
-                rds_message_put(ic->i_rm);
+                struct rds_message *rm;
-                ic->i_rm = NULL;
+                rm = container_of(ic->i_data_op, struct rds_message, data);
+                rds_message_put(rm);
+                ic->i_data_op = NULL;
        }
        /* Clear the ACK state */
@@ -690,12 +748,19 @@ int rds_ib_conn_alloc(struct rds_connection *conn, gfp_t gfp)
 {
        struct rds_ib_connection *ic;
        unsigned long flags;
+        int ret;
        /* XXX too lazy? */
        ic = kzalloc(sizeof(struct rds_ib_connection), GFP_KERNEL);
-        if (ic == NULL)
+        if (!ic)
                return -ENOMEM;
+        ret = rds_ib_recv_alloc_caches(ic);
+        if (ret) {
+                kfree(ic);
+                return ret;
+        }
        INIT_LIST_HEAD(&ic->ib_node);
        tasklet_init(&ic->i_recv_tasklet, rds_ib_recv_tasklet_fn,
                     (unsigned long) ic);
@@ -703,6 +768,7 @@ int rds_ib_conn_alloc(struct rds_connection *conn, gfp_t gfp)
 #ifndef KERNEL_HAS_ATOMIC64
        spin_lock_init(&ic->i_ack_lock);
 #endif
+        atomic_set(&ic->i_signaled_sends, 0);
        /*
         * rds_ib_conn_shutdown() waits for these to be emptied so they
@@ -744,6 +810,8 @@ void rds_ib_conn_free(void *arg)
        list_del(&ic->ib_node);
        spin_unlock_irq(lock_ptr);
+        rds_ib_recv_free_caches(ic);
        kfree(ic);
 }
diff --git a/net/rds/ib_rdma.c b/net/rds/ib_rdma.c
index a54cd63f9e35..b5a88415a18e 100644
--- a/net/rds/ib_rdma.c
+++ b/net/rds/ib_rdma.c
@@ -32,11 +32,16 @@
 */
 #include <linux/kernel.h>
 #include <linux/slab.h>
+#include <linux/rculist.h>
 #include "rds.h"
-#include "rdma.h"
 #include "ib.h"
+#include "xlist.h"
+struct workqueue_struct *rds_ib_fmr_wq;
+static DEFINE_PER_CPU(unsigned long, clean_list_grace);
+#define CLEAN_LIST_BUSY_BIT 0
 /*
 * This is stored as mr->r_trans_private.
@@ -45,7 +50,11 @@ struct rds_ib_mr {
        struct rds_ib_device    *device;
        struct rds_ib_mr_pool   *pool;
        struct ib_fmr           *fmr;
-        struct list_head        list;
+        struct xlist_head       xlist;
+        /* unmap_list is for freeing */
+        struct list_head        unmap_list;
        unsigned int            remap_count;
        struct scatterlist      *sg;
@@ -59,14 +68,16 @@ struct rds_ib_mr {
 */
 struct rds_ib_mr_pool {
        struct mutex            flush_lock;             /* serialize fmr invalidate */
-        struct work_struct      flush_worker;           /* flush worker */
+        struct delayed_work     flush_worker;           /* flush worker */
-        spinlock_t              list_lock;              /* protect variables below */
        atomic_t                item_count;             /* total # of MRs */
        atomic_t                dirty_count;            /* # dirty of MRs */
-        struct list_head        drop_list;              /* MRs that have reached their max_maps limit */
-        struct list_head        free_list;              /* unused MRs */
+        struct xlist_head       drop_list;              /* MRs that have reached their max_maps limit */
-        struct list_head        clean_list;             /* unused & unamapped MRs */
+        struct xlist_head       free_list;              /* unused MRs */
+        struct xlist_head       clean_list;             /* global unused & unamapped MRs */
+        wait_queue_head_t       flush_wait;
        atomic_t                free_pinned;            /* memory pinned by free MRs */
        unsigned long           max_items;
        unsigned long           max_items_soft;
@@ -74,7 +85,7 @@ struct rds_ib_mr_pool {
        struct ib_fmr_attr      fmr_attr;
 };
-static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all);
+static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all, struct rds_ib_mr **);
 static void rds_ib_teardown_mr(struct rds_ib_mr *ibmr);
 static void rds_ib_mr_pool_flush_worker(struct work_struct *work);
@@ -83,16 +94,17 @@ static struct rds_ib_device *rds_ib_get_device(__be32 ipaddr)
        struct rds_ib_device *rds_ibdev;
        struct rds_ib_ipaddr *i_ipaddr;
-        list_for_each_entry(rds_ibdev, &rds_ib_devices, list) {
+        rcu_read_lock();
-                spin_lock_irq(&rds_ibdev->spinlock);
+        list_for_each_entry_rcu(rds_ibdev, &rds_ib_devices, list) {
-                list_for_each_entry(i_ipaddr, &rds_ibdev->ipaddr_list, list) {
+                list_for_each_entry_rcu(i_ipaddr, &rds_ibdev->ipaddr_list, list) {
                        if (i_ipaddr->ipaddr == ipaddr) {
-                                spin_unlock_irq(&rds_ibdev->spinlock);
+                                atomic_inc(&rds_ibdev->refcount);
+                                rcu_read_unlock();
                                return rds_ibdev;
                        }
                }
-                spin_unlock_irq(&rds_ibdev->spinlock);
        }
+        rcu_read_unlock();
        return NULL;
 }
@@ -108,7 +120,7 @@ static int rds_ib_add_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr)
        i_ipaddr->ipaddr = ipaddr;
        spin_lock_irq(&rds_ibdev->spinlock);
-        list_add_tail(&i_ipaddr->list, &rds_ibdev->ipaddr_list);
+        list_add_tail_rcu(&i_ipaddr->list, &rds_ibdev->ipaddr_list);
        spin_unlock_irq(&rds_ibdev->spinlock);
        return 0;
@@ -116,17 +128,24 @@ static int rds_ib_add_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr)
 static void rds_ib_remove_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr)
 {
-        struct rds_ib_ipaddr *i_ipaddr, *next;
+        struct rds_ib_ipaddr *i_ipaddr;
+        struct rds_ib_ipaddr *to_free = NULL;
        spin_lock_irq(&rds_ibdev->spinlock);
-        list_for_each_entry_safe(i_ipaddr, next, &rds_ibdev->ipaddr_list, list) {
+        list_for_each_entry_rcu(i_ipaddr, &rds_ibdev->ipaddr_list, list) {
                if (i_ipaddr->ipaddr == ipaddr) {
-                        list_del(&i_ipaddr->list);
+                        list_del_rcu(&i_ipaddr->list);
-                        kfree(i_ipaddr);
+                        to_free = i_ipaddr;
                        break;
                }
        }
        spin_unlock_irq(&rds_ibdev->spinlock);
+        if (to_free) {
+                synchronize_rcu();
+                kfree(to_free);
+        }
 }
 int rds_ib_update_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr)
@@ -134,8 +153,10 @@ int rds_ib_update_ipaddr(struct rds_ib_device *rds_ibdev, __be32 ipaddr)
        struct rds_ib_device *rds_ibdev_old;
        rds_ibdev_old = rds_ib_get_device(ipaddr);
-        if (rds_ibdev_old)
+        if (rds_ibdev_old) {
                rds_ib_remove_ipaddr(rds_ibdev_old, ipaddr);
+                rds_ib_dev_put(rds_ibdev_old);
+        }
        return rds_ib_add_ipaddr(rds_ibdev, ipaddr);
 }
@@ -150,12 +171,13 @@ void rds_ib_add_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *con
        BUG_ON(list_empty(&ic->ib_node));
        list_del(&ic->ib_node);
-        spin_lock_irq(&rds_ibdev->spinlock);
+        spin_lock(&rds_ibdev->spinlock);
        list_add_tail(&ic->ib_node, &rds_ibdev->conn_list);
-        spin_unlock_irq(&rds_ibdev->spinlock);
+        spin_unlock(&rds_ibdev->spinlock);
        spin_unlock_irq(&ib_nodev_conns_lock);
        ic->rds_ibdev = rds_ibdev;
+        atomic_inc(&rds_ibdev->refcount);
 }
 void rds_ib_remove_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *conn)
@@ -175,18 +197,18 @@ void rds_ib_remove_conn(struct rds_ib_device *rds_ibdev, struct rds_connection *
        spin_unlock(&ib_nodev_conns_lock);
        ic->rds_ibdev = NULL;
+        rds_ib_dev_put(rds_ibdev);
 }
-void __rds_ib_destroy_conns(struct list_head *list, spinlock_t *list_lock)
+void rds_ib_destroy_nodev_conns(void)
 {
        struct rds_ib_connection *ic, *_ic;
        LIST_HEAD(tmp_list);
        /* avoid calling conn_destroy with irqs off */
-        spin_lock_irq(list_lock);
+        spin_lock_irq(&ib_nodev_conns_lock);
-        list_splice(list, &tmp_list);
+        list_splice(&ib_nodev_conns, &tmp_list);
-        INIT_LIST_HEAD(list);
+        spin_unlock_irq(&ib_nodev_conns_lock);
-        spin_unlock_irq(list_lock);
        list_for_each_entry_safe(ic, _ic, &tmp_list, ib_node)
                rds_conn_destroy(ic->conn);
@@ -200,12 +222,12 @@ struct rds_ib_mr_pool *rds_ib_create_mr_pool(struct rds_ib_device *rds_ibdev)
        if (!pool)
                return ERR_PTR(-ENOMEM);
-        INIT_LIST_HEAD(&pool->free_list);
+        INIT_XLIST_HEAD(&pool->free_list);
-        INIT_LIST_HEAD(&pool->drop_list);
+        INIT_XLIST_HEAD(&pool->drop_list);
-        INIT_LIST_HEAD(&pool->clean_list);
+        INIT_XLIST_HEAD(&pool->clean_list);
        mutex_init(&pool->flush_lock);
-        spin_lock_init(&pool->list_lock);
+        init_waitqueue_head(&pool->flush_wait);
-        INIT_WORK(&pool->flush_worker, rds_ib_mr_pool_flush_worker);
+        INIT_DELAYED_WORK(&pool->flush_worker, rds_ib_mr_pool_flush_worker);
        pool->fmr_attr.max_pages = fmr_message_size;
        pool->fmr_attr.max_maps = rds_ibdev->fmr_max_remaps;
@@ -233,34 +255,60 @@ void rds_ib_get_mr_info(struct rds_ib_device *rds_ibdev, struct rds_info_rdma_co
 void rds_ib_destroy_mr_pool(struct rds_ib_mr_pool *pool)
 {
-        flush_workqueue(rds_wq);
+        cancel_delayed_work_sync(&pool->flush_worker);
-        rds_ib_flush_mr_pool(pool, 1);
+        rds_ib_flush_mr_pool(pool, 1, NULL);
        WARN_ON(atomic_read(&pool->item_count));
        WARN_ON(atomic_read(&pool->free_pinned));
        kfree(pool);
 }
+static void refill_local(struct rds_ib_mr_pool *pool, struct xlist_head *xl,
+                         struct rds_ib_mr **ibmr_ret)
+{
+        struct xlist_head *ibmr_xl;
+        ibmr_xl = xlist_del_head_fast(xl);
+        *ibmr_ret = list_entry(ibmr_xl, struct rds_ib_mr, xlist);
+}
 static inline struct rds_ib_mr *rds_ib_reuse_fmr(struct rds_ib_mr_pool *pool)
 {
        struct rds_ib_mr *ibmr = NULL;
-        unsigned long flags;
+        struct xlist_head *ret;
+        unsigned long *flag;
-        spin_lock_irqsave(&pool->list_lock, flags);
+        preempt_disable();
-        if (!list_empty(&pool->clean_list)) {
+        flag = &__get_cpu_var(clean_list_grace);
-                ibmr = list_entry(pool->clean_list.next, struct rds_ib_mr, list);
+        set_bit(CLEAN_LIST_BUSY_BIT, flag);
-                list_del_init(&ibmr->list);
+        ret = xlist_del_head(&pool->clean_list);
-        }
+        if (ret)
-        spin_unlock_irqrestore(&pool->list_lock, flags);
+                ibmr = list_entry(ret, struct rds_ib_mr, xlist);
+        clear_bit(CLEAN_LIST_BUSY_BIT, flag);
+        preempt_enable();
        return ibmr;
 }
+static inline void wait_clean_list_grace(void)
+{
+        int cpu;
+        unsigned long *flag;
+        for_each_online_cpu(cpu) {
+                flag = &per_cpu(clean_list_grace, cpu);
+                while (test_bit(CLEAN_LIST_BUSY_BIT, flag))
+                        cpu_relax();
+        }
+}
 static struct rds_ib_mr *rds_ib_alloc_fmr(struct rds_ib_device *rds_ibdev)
 {
        struct rds_ib_mr_pool *pool = rds_ibdev->mr_pool;
        struct rds_ib_mr *ibmr = NULL;
        int err = 0, iter = 0;
+        if (atomic_read(&pool->dirty_count) >= pool->max_items / 10)
+                queue_delayed_work(rds_ib_fmr_wq, &pool->flush_worker, 10);
        while (1) {
                ibmr = rds_ib_reuse_fmr(pool);
                if (ibmr)
@@ -287,19 +335,24 @@ static struct rds_ib_mr *rds_ib_alloc_fmr(struct rds_ib_device *rds_ibdev)
                /* We do have some empty MRs. Flush them out. */
                rds_ib_stats_inc(s_ib_rdma_mr_pool_wait);
-                rds_ib_flush_mr_pool(pool, 0);
+                rds_ib_flush_mr_pool(pool, 0, &ibmr);
+                if (ibmr)
+                        return ibmr;
        }
-        ibmr = kzalloc(sizeof(*ibmr), GFP_KERNEL);
+        ibmr = kzalloc_node(sizeof(*ibmr), GFP_KERNEL, rdsibdev_to_node(rds_ibdev));
        if (!ibmr) {
                err = -ENOMEM;
                goto out_no_cigar;
        }
+        memset(ibmr, 0, sizeof(*ibmr));
        ibmr->fmr = ib_alloc_fmr(rds_ibdev->pd,
                        (IB_ACCESS_LOCAL_WRITE |
                         IB_ACCESS_REMOTE_READ |
-                         IB_ACCESS_REMOTE_WRITE),
+                         IB_ACCESS_REMOTE_WRITE|
+                         IB_ACCESS_REMOTE_ATOMIC),
                        &pool->fmr_attr);
        if (IS_ERR(ibmr->fmr)) {
                err = PTR_ERR(ibmr->fmr);
@@ -367,7 +420,8 @@ static int rds_ib_map_fmr(struct rds_ib_device *rds_ibdev, struct rds_ib_mr *ibm
        if (page_cnt > fmr_message_size)
                return -EINVAL;
-        dma_pages = kmalloc(sizeof(u64) * page_cnt, GFP_ATOMIC);
+        dma_pages = kmalloc_node(sizeof(u64) * page_cnt, GFP_ATOMIC,
+                                 rdsibdev_to_node(rds_ibdev));
        if (!dma_pages)
                return -ENOMEM;
@@ -441,7 +495,7 @@ static void __rds_ib_teardown_mr(struct rds_ib_mr *ibmr)
                        /* FIXME we need a way to tell a r/w MR
                         * from a r/o MR */
-                        BUG_ON(in_interrupt());
+                        BUG_ON(irqs_disabled());
                        set_page_dirty(page);
                        put_page(page);
                }
@@ -477,33 +531,109 @@ static inline unsigned int rds_ib_flush_goal(struct rds_ib_mr_pool *pool, int fr
 }
 /*
+ * given an xlist of mrs, put them all into the list_head for more processing
+ */
+static void xlist_append_to_list(struct xlist_head *xlist, struct list_head *list)
+{
+        struct rds_ib_mr *ibmr;
+        struct xlist_head splice;
+        struct xlist_head *cur;
+        struct xlist_head *next;
+        splice.next = NULL;
+        xlist_splice(xlist, &splice);
+        cur = splice.next;
+        while (cur) {
+                next = cur->next;
+                ibmr = list_entry(cur, struct rds_ib_mr, xlist);
+                list_add_tail(&ibmr->unmap_list, list);
+                cur = next;
+        }
+}
+/*
+ * this takes a list head of mrs and turns it into an xlist of clusters.
+ * each cluster has an xlist of MR_CLUSTER_SIZE mrs that are ready for
+ * reuse.
+ */
+static void list_append_to_xlist(struct rds_ib_mr_pool *pool,
+                                struct list_head *list, struct xlist_head *xlist,
+                                struct xlist_head **tail_ret)
+{
+        struct rds_ib_mr *ibmr;
+        struct xlist_head *cur_mr = xlist;
+        struct xlist_head *tail_mr = NULL;
+        list_for_each_entry(ibmr, list, unmap_list) {
+                tail_mr = &ibmr->xlist;
+                tail_mr->next = NULL;
+                cur_mr->next = tail_mr;
+                cur_mr = tail_mr;
+        }
+        *tail_ret = tail_mr;
+}
+/*
 * Flush our pool of MRs.
 * At a minimum, all currently unused MRs are unmapped.
 * If the number of MRs allocated exceeds the limit, we also try
 * to free as many MRs as needed to get back to this limit.
 */
-static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all)
+static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool,
+                                int free_all, struct rds_ib_mr **ibmr_ret)
 {
        struct rds_ib_mr *ibmr, *next;
+        struct xlist_head clean_xlist;
+        struct xlist_head *clean_tail;
        LIST_HEAD(unmap_list);
        LIST_HEAD(fmr_list);
        unsigned long unpinned = 0;
-        unsigned long flags;
        unsigned int nfreed = 0, ncleaned = 0, free_goal;
        int ret = 0;
        rds_ib_stats_inc(s_ib_rdma_mr_pool_flush);
-        mutex_lock(&pool->flush_lock);
+        if (ibmr_ret) {
+                DEFINE_WAIT(wait);
+                while(!mutex_trylock(&pool->flush_lock)) {
+                        ibmr = rds_ib_reuse_fmr(pool);
+                        if (ibmr) {
+                                *ibmr_ret = ibmr;
+                                finish_wait(&pool->flush_wait, &wait);
+                                goto out_nolock;
+                        }
+                        prepare_to_wait(&pool->flush_wait, &wait,
+                                        TASK_UNINTERRUPTIBLE);
+                        if (xlist_empty(&pool->clean_list))
+                                schedule();
+                        ibmr = rds_ib_reuse_fmr(pool);
+                        if (ibmr) {
+                                *ibmr_ret = ibmr;
+                                finish_wait(&pool->flush_wait, &wait);
+                                goto out_nolock;
+                        }
+                }
+                finish_wait(&pool->flush_wait, &wait);
+        } else
+                mutex_lock(&pool->flush_lock);
+        if (ibmr_ret) {
+                ibmr = rds_ib_reuse_fmr(pool);
+                if (ibmr) {
+                        *ibmr_ret = ibmr;
+                        goto out;
+                }
+        }
-        spin_lock_irqsave(&pool->list_lock, flags);
        /* Get the list of all MRs to be dropped. Ordering matters -
-         * we want to put drop_list ahead of free_list. */
+         * we want to put drop_list ahead of free_list.
-        list_splice_init(&pool->free_list, &unmap_list);
+         */
-        list_splice_init(&pool->drop_list, &unmap_list);
+        xlist_append_to_list(&pool->drop_list, &unmap_list);
+        xlist_append_to_list(&pool->free_list, &unmap_list);
        if (free_all)
-                list_splice_init(&pool->clean_list, &unmap_list);
+                xlist_append_to_list(&pool->clean_list, &unmap_list);
-        spin_unlock_irqrestore(&pool->list_lock, flags);
        free_goal = rds_ib_flush_goal(pool, free_all);
@@ -511,19 +641,20 @@ static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all)
                goto out;
        /* String all ib_mr's onto one list and hand them to ib_unmap_fmr */
-        list_for_each_entry(ibmr, &unmap_list, list)
+        list_for_each_entry(ibmr, &unmap_list, unmap_list)
                list_add(&ibmr->fmr->list, &fmr_list);
        ret = ib_unmap_fmr(&fmr_list);
        if (ret)
                printk(KERN_WARNING "RDS/IB: ib_unmap_fmr failed (err=%d)\n", ret);
        /* Now we can destroy the DMA mapping and unpin any pages */
-        list_for_each_entry_safe(ibmr, next, &unmap_list, list) {
+        list_for_each_entry_safe(ibmr, next, &unmap_list, unmap_list) {
                unpinned += ibmr->sg_len;
                __rds_ib_teardown_mr(ibmr);
                if (nfreed < free_goal || ibmr->remap_count >= pool->fmr_attr.max_maps) {
                        rds_ib_stats_inc(s_ib_rdma_mr_free);
-                        list_del(&ibmr->list);
+                        list_del(&ibmr->unmap_list);
                        ib_dealloc_fmr(ibmr->fmr);
                        kfree(ibmr);
                        nfreed++;
@@ -531,9 +662,27 @@ static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all)
                ncleaned++;
        }
-        spin_lock_irqsave(&pool->list_lock, flags);
+        if (!list_empty(&unmap_list)) {
-        list_splice(&unmap_list, &pool->clean_list);
+                /* we have to make sure that none of the things we're about
-        spin_unlock_irqrestore(&pool->list_lock, flags);
+                 * to put on the clean list would race with other cpus trying
+                 * to pull items off.  The xlist would explode if we managed to
+                 * remove something from the clean list and then add it back again
+                 * while another CPU was spinning on that same item in xlist_del_head.
+                 *
+                 * This is pretty unlikely, but just in case  wait for an xlist grace period
+                 * here before adding anything back into the clean list.
+                 */
+                wait_clean_list_grace();
+                list_append_to_xlist(pool, &unmap_list, &clean_xlist, &clean_tail);
+                if (ibmr_ret)
+                        refill_local(pool, &clean_xlist, ibmr_ret);
+                /* refill_local may have emptied our list */
+                if (!xlist_empty(&clean_xlist))
+                        xlist_add(clean_xlist.next, clean_tail, &pool->clean_list);
+        }
        atomic_sub(unpinned, &pool->free_pinned);
        atomic_sub(ncleaned, &pool->dirty_count);
@@ -541,14 +690,35 @@ static int rds_ib_flush_mr_pool(struct rds_ib_mr_pool *pool, int free_all)
 out:
        mutex_unlock(&pool->flush_lock);
+        if (waitqueue_active(&pool->flush_wait))
+                wake_up(&pool->flush_wait);
+out_nolock:
        return ret;
 }
+int rds_ib_fmr_init(void)
+{
+        rds_ib_fmr_wq = create_workqueue("rds_fmr_flushd");
+        if (!rds_ib_fmr_wq)
+                return -ENOMEM;
+        return 0;
+}
+/*
+ * By the time this is called all the IB devices should have been torn down and
+ * had their pools freed.  As each pool is freed its work struct is waited on,
+ * so the pool flushing work queue should be idle by the time we get here.
+ */
+void rds_ib_fmr_exit(void)
+{
+        destroy_workqueue(rds_ib_fmr_wq);
+}
 static void rds_ib_mr_pool_flush_worker(struct work_struct *work)
 {
-        struct rds_ib_mr_pool *pool = container_of(work, struct rds_ib_mr_pool, flush_worker);
+        struct rds_ib_mr_pool *pool = container_of(work, struct rds_ib_mr_pool, flush_worker.work);
-        rds_ib_flush_mr_pool(pool, 0);
+        rds_ib_flush_mr_pool(pool, 0, NULL);
 }
 void rds_ib_free_mr(void *trans_private, int invalidate)
@@ -556,47 +726,49 @@ void rds_ib_free_mr(void *trans_private, int invalidate)
        struct rds_ib_mr *ibmr = trans_private;
        struct rds_ib_device *rds_ibdev = ibmr->device;
        struct rds_ib_mr_pool *pool = rds_ibdev->mr_pool;
-        unsigned long flags;
        rdsdebug("RDS/IB: free_mr nents %u\n", ibmr->sg_len);
        /* Return it to the pool's free list */
-        spin_lock_irqsave(&pool->list_lock, flags);
        if (ibmr->remap_count >= pool->fmr_attr.max_maps)
-                list_add(&ibmr->list, &pool->drop_list);
+                xlist_add(&ibmr->xlist, &ibmr->xlist, &pool->drop_list);
        else
-                list_add(&ibmr->list, &pool->free_list);
+                xlist_add(&ibmr->xlist, &ibmr->xlist, &pool->free_list);
        atomic_add(ibmr->sg_len, &pool->free_pinned);
        atomic_inc(&pool->dirty_count);
-        spin_unlock_irqrestore(&pool->list_lock, flags);
        /* If we've pinned too many pages, request a flush */
        if (atomic_read(&pool->free_pinned) >= pool->max_free_pinned ||
            atomic_read(&pool->dirty_count) >= pool->max_items / 10)
-                queue_work(rds_wq, &pool->flush_worker);
+                queue_delayed_work(rds_ib_fmr_wq, &pool->flush_worker, 10);
        if (invalidate) {
                if (likely(!in_interrupt())) {
-                        rds_ib_flush_mr_pool(pool, 0);
+                        rds_ib_flush_mr_pool(pool, 0, NULL);
                } else {
                        /* We get here if the user created a MR marked
                         * as use_once and invalidate at the same time. */
-                        queue_work(rds_wq, &pool->flush_worker);
+                        queue_delayed_work(rds_ib_fmr_wq,
+                                           &pool->flush_worker, 10);
                }
        }
+        rds_ib_dev_put(rds_ibdev);
 }
 void rds_ib_flush_mrs(void)
 {
        struct rds_ib_device *rds_ibdev;
+        down_read(&rds_ib_devices_lock);
        list_for_each_entry(rds_ibdev, &rds_ib_devices, list) {
                struct rds_ib_mr_pool *pool = rds_ibdev->mr_pool;
                if (pool)
-                        rds_ib_flush_mr_pool(pool, 0);
+                        rds_ib_flush_mr_pool(pool, 0, NULL);
        }
+        up_read(&rds_ib_devices_lock);
 }
 void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
@@ -628,6 +800,7 @@ void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
                printk(KERN_WARNING "RDS/IB: map_fmr failed (errno=%d)\n", ret);
        ibmr->device = rds_ibdev;
+        rds_ibdev = NULL;
 out:
        if (ret) {
@@ -635,5 +808,8 @@ void *rds_ib_get_mr(struct scatterlist *sg, unsigned long nents,
                        rds_ib_free_mr(ibmr, 0);
                ibmr = ERR_PTR(ret);
        }
+        if (rds_ibdev)
+                rds_ib_dev_put(rds_ibdev);
        return ibmr;
 }
diff --git a/net/rds/ib_recv.c b/net/rds/ib_recv.c
index c74e9904a6b2..e29e0ca32f74 100644
--- a/net/rds/ib_recv.c
+++ b/net/rds/ib_recv.c
@@ -43,42 +43,6 @@ static struct kmem_cache *rds_ib_incoming_slab;
 static struct kmem_cache *rds_ib_frag_slab;
 static atomic_t rds_ib_allocation = ATOMIC_INIT(0);
-static void rds_ib_frag_drop_page(struct rds_page_frag *frag)
-{
-        rdsdebug("frag %p page %p\n", frag, frag->f_page);
-        __free_page(frag->f_page);
-        frag->f_page = NULL;
-}
-static void rds_ib_frag_free(struct rds_page_frag *frag)
-{
-        rdsdebug("frag %p page %p\n", frag, frag->f_page);
-        BUG_ON(frag->f_page != NULL);
-        kmem_cache_free(rds_ib_frag_slab, frag);
-}
-/*
- * We map a page at a time.  Its fragments are posted in order.  This
- * is called in fragment order as the fragments get send completion events.
- * Only the last frag in the page performs the unmapping.
- *
- * It's OK for ring cleanup to call this in whatever order it likes because
- * DMA is not in flight and so we can unmap while other ring entries still
- * hold page references in their frags.
- */
-static void rds_ib_recv_unmap_page(struct rds_ib_connection *ic,
-                                   struct rds_ib_recv_work *recv)
-{
-        struct rds_page_frag *frag = recv->r_frag;
-        rdsdebug("recv %p frag %p page %p\n", recv, frag, frag->f_page);
-        if (frag->f_mapped)
-                ib_dma_unmap_page(ic->i_cm_id->device,
-                               frag->f_mapped,
-                               RDS_FRAG_SIZE, DMA_FROM_DEVICE);
-        frag->f_mapped = 0;
-}
 void rds_ib_recv_init_ring(struct rds_ib_connection *ic)
 {
        struct rds_ib_recv_work *recv;
@@ -95,16 +59,161 @@ void rds_ib_recv_init_ring(struct rds_ib_connection *ic)
                recv->r_wr.sg_list = recv->r_sge;
                recv->r_wr.num_sge = RDS_IB_RECV_SGE;
-                sge = rds_ib_data_sge(ic, recv->r_sge);
+                sge = &recv->r_sge[0];
+                sge->addr = ic->i_recv_hdrs_dma + (i * sizeof(struct rds_header));
+                sge->length = sizeof(struct rds_header);
+                sge->lkey = ic->i_mr->lkey;
+                sge = &recv->r_sge[1];
                sge->addr = 0;
                sge->length = RDS_FRAG_SIZE;
                sge->lkey = ic->i_mr->lkey;
+        }
+}
-                sge = rds_ib_header_sge(ic, recv->r_sge);
+/*
-                sge->addr = ic->i_recv_hdrs_dma + (i * sizeof(struct rds_header));
+ * The entire 'from' list, including the from element itself, is put on
-                sge->length = sizeof(struct rds_header);
+ * to the tail of the 'to' list.
-                sge->lkey = ic->i_mr->lkey;
+ */
+static void list_splice_entire_tail(struct list_head *from,
+                                    struct list_head *to)
+{
+        struct list_head *from_last = from->prev;
+        list_splice_tail(from_last, to);
+        list_add_tail(from_last, to);
+}
+static void rds_ib_cache_xfer_to_ready(struct rds_ib_refill_cache *cache)
+{
+        struct list_head *tmp;
+        tmp = xchg(&cache->xfer, NULL);
+        if (tmp) {
+                if (cache->ready)
+                        list_splice_entire_tail(tmp, cache->ready);
+                else
+                        cache->ready = tmp;
+        }
+}
+static int rds_ib_recv_alloc_cache(struct rds_ib_refill_cache *cache)
+{
+        struct rds_ib_cache_head *head;
+        int cpu;
+        cache->percpu = alloc_percpu(struct rds_ib_cache_head);
+        if (!cache->percpu)
+               return -ENOMEM;
+        for_each_possible_cpu(cpu) {
+                head = per_cpu_ptr(cache->percpu, cpu);
+                head->first = NULL;
+                head->count = 0;
+        }
+        cache->xfer = NULL;
+        cache->ready = NULL;
+        return 0;
+}
+int rds_ib_recv_alloc_caches(struct rds_ib_connection *ic)
+{
+        int ret;
+        ret = rds_ib_recv_alloc_cache(&ic->i_cache_incs);
+        if (!ret) {
+                ret = rds_ib_recv_alloc_cache(&ic->i_cache_frags);
+                if (ret)
+                        free_percpu(ic->i_cache_incs.percpu);
        }
+        return ret;
+}
+static void rds_ib_cache_splice_all_lists(struct rds_ib_refill_cache *cache,
+                                          struct list_head *caller_list)
+{
+        struct rds_ib_cache_head *head;
+        int cpu;
+        for_each_possible_cpu(cpu) {
+                head = per_cpu_ptr(cache->percpu, cpu);
+                if (head->first) {
+                        list_splice_entire_tail(head->first, caller_list);
+                        head->first = NULL;
+                }
+        }
+        if (cache->ready) {
+                list_splice_entire_tail(cache->ready, caller_list);
+                cache->ready = NULL;
+        }
+}
+void rds_ib_recv_free_caches(struct rds_ib_connection *ic)
+{
+        struct rds_ib_incoming *inc;
+        struct rds_ib_incoming *inc_tmp;
+        struct rds_page_frag *frag;
+        struct rds_page_frag *frag_tmp;
+        LIST_HEAD(list);
+        rds_ib_cache_xfer_to_ready(&ic->i_cache_incs);
+        rds_ib_cache_splice_all_lists(&ic->i_cache_incs, &list);
+        free_percpu(ic->i_cache_incs.percpu);
+        list_for_each_entry_safe(inc, inc_tmp, &list, ii_cache_entry) {
+                list_del(&inc->ii_cache_entry);
+                WARN_ON(!list_empty(&inc->ii_frags));
+                kmem_cache_free(rds_ib_incoming_slab, inc);
+        }
+        rds_ib_cache_xfer_to_ready(&ic->i_cache_frags);
+        rds_ib_cache_splice_all_lists(&ic->i_cache_frags, &list);
+        free_percpu(ic->i_cache_frags.percpu);
+        list_for_each_entry_safe(frag, frag_tmp, &list, f_cache_entry) {
+                list_del(&frag->f_cache_entry);
+                WARN_ON(!list_empty(&frag->f_item));
+                kmem_cache_free(rds_ib_frag_slab, frag);
+        }
+}
+/* fwd decl */
+static void rds_ib_recv_cache_put(struct list_head *new_item,
+                                  struct rds_ib_refill_cache *cache);
+static struct list_head *rds_ib_recv_cache_get(struct rds_ib_refill_cache *cache);
+/* Recycle frag and attached recv buffer f_sg */
+static void rds_ib_frag_free(struct rds_ib_connection *ic,
+                             struct rds_page_frag *frag)
+{
+        rdsdebug("frag %p page %p\n", frag, sg_page(&frag->f_sg));
+        rds_ib_recv_cache_put(&frag->f_cache_entry, &ic->i_cache_frags);
+}
+/* Recycle inc after freeing attached frags */
+void rds_ib_inc_free(struct rds_incoming *inc)
+{
+        struct rds_ib_incoming *ibinc;
+        struct rds_page_frag *frag;
+        struct rds_page_frag *pos;
+        struct rds_ib_connection *ic = inc->i_conn->c_transport_data;
+        ibinc = container_of(inc, struct rds_ib_incoming, ii_inc);
+        /* Free attached frags */
+        list_for_each_entry_safe(frag, pos, &ibinc->ii_frags, f_item) {
+                list_del_init(&frag->f_item);
+                rds_ib_frag_free(ic, frag);
+        }
+        BUG_ON(!list_empty(&ibinc->ii_frags));
+        rdsdebug("freeing ibinc %p inc %p\n", ibinc, inc);
+        rds_ib_recv_cache_put(&ibinc->ii_cache_entry, &ic->i_cache_incs);
 }
 static void rds_ib_recv_clear_one(struct rds_ib_connection *ic,
@@ -115,10 +224,8 @@ static void rds_ib_recv_clear_one(struct rds_ib_connection *ic,
                recv->r_ibinc = NULL;
        }
        if (recv->r_frag) {
-                rds_ib_recv_unmap_page(ic, recv);
+                ib_dma_unmap_sg(ic->i_cm_id->device, &recv->r_frag->f_sg, 1, DMA_FROM_DEVICE);
-                if (recv->r_frag->f_page)
+                rds_ib_frag_free(ic, recv->r_frag);
-                        rds_ib_frag_drop_page(recv->r_frag);
-                rds_ib_frag_free(recv->r_frag);
                recv->r_frag = NULL;
        }
 }
@@ -129,84 +236,111 @@ void rds_ib_recv_clear_ring(struct rds_ib_connection *ic)
        for (i = 0; i < ic->i_recv_ring.w_nr; i++)
                rds_ib_recv_clear_one(ic, &ic->i_recvs[i]);
-        if (ic->i_frag.f_page)
-                rds_ib_frag_drop_page(&ic->i_frag);
 }
-static int rds_ib_recv_refill_one(struct rds_connection *conn,
+static struct rds_ib_incoming *rds_ib_refill_one_inc(struct rds_ib_connection *ic,
-                                  struct rds_ib_recv_work *recv,
+                                                     gfp_t slab_mask)
-                                  gfp_t kptr_gfp, gfp_t page_gfp)
 {
-        struct rds_ib_connection *ic = conn->c_transport_data;
+        struct rds_ib_incoming *ibinc;
-        dma_addr_t dma_addr;
+        struct list_head *cache_item;
-        struct ib_sge *sge;
+        int avail_allocs;
-        int ret = -ENOMEM;
-        if (recv->r_ibinc == NULL) {
+        cache_item = rds_ib_recv_cache_get(&ic->i_cache_incs);
-                if (!atomic_add_unless(&rds_ib_allocation, 1, rds_ib_sysctl_max_recv_allocation)) {
+        if (cache_item) {
+                ibinc = container_of(cache_item, struct rds_ib_incoming, ii_cache_entry);
+        } else {
+                avail_allocs = atomic_add_unless(&rds_ib_allocation,
+                                                 1, rds_ib_sysctl_max_recv_allocation);
+                if (!avail_allocs) {
                        rds_ib_stats_inc(s_ib_rx_alloc_limit);
-                        goto out;
+                        return NULL;
                }
-                recv->r_ibinc = kmem_cache_alloc(rds_ib_incoming_slab,
+                ibinc = kmem_cache_alloc(rds_ib_incoming_slab, slab_mask);
-                                                 kptr_gfp);
+                if (!ibinc) {
-                if (recv->r_ibinc == NULL) {
                        atomic_dec(&rds_ib_allocation);
-                        goto out;
+                        return NULL;
                }
-                INIT_LIST_HEAD(&recv->r_ibinc->ii_frags);
-                rds_inc_init(&recv->r_ibinc->ii_inc, conn, conn->c_faddr);
        }
+        INIT_LIST_HEAD(&ibinc->ii_frags);
+        rds_inc_init(&ibinc->ii_inc, ic->conn, ic->conn->c_faddr);
-        if (recv->r_frag == NULL) {
+        return ibinc;
-                recv->r_frag = kmem_cache_alloc(rds_ib_frag_slab, kptr_gfp);
+}
-                if (recv->r_frag == NULL)
-                        goto out;
+static struct rds_page_frag *rds_ib_refill_one_frag(struct rds_ib_connection *ic,
-                INIT_LIST_HEAD(&recv->r_frag->f_item);
+                                                    gfp_t slab_mask, gfp_t page_mask)
-                recv->r_frag->f_page = NULL;
+{
+        struct rds_page_frag *frag;
+        struct list_head *cache_item;
+        int ret;
+        cache_item = rds_ib_recv_cache_get(&ic->i_cache_frags);
+        if (cache_item) {
+                frag = container_of(cache_item, struct rds_page_frag, f_cache_entry);
+        } else {
+                frag = kmem_cache_alloc(rds_ib_frag_slab, slab_mask);
+                if (!frag)
+                        return NULL;
+                sg_init_table(&frag->f_sg, 1);
+                ret = rds_page_remainder_alloc(&frag->f_sg,
+                                               RDS_FRAG_SIZE, page_mask);
+                if (ret) {
+                        kmem_cache_free(rds_ib_frag_slab, frag);
+                        return NULL;
+                }
        }
-        if (ic->i_frag.f_page == NULL) {
+        INIT_LIST_HEAD(&frag->f_item);
-                ic->i_frag.f_page = alloc_page(page_gfp);
-                if (ic->i_frag.f_page == NULL)
+        return frag;
-                        goto out;
+}
-                ic->i_frag.f_offset = 0;
+static int rds_ib_recv_refill_one(struct rds_connection *conn,
+                                  struct rds_ib_recv_work *recv, int prefill)
+{
+        struct rds_ib_connection *ic = conn->c_transport_data;
+        struct ib_sge *sge;
+        int ret = -ENOMEM;
+        gfp_t slab_mask = GFP_NOWAIT;
+        gfp_t page_mask = GFP_NOWAIT;
+        if (prefill) {
+                slab_mask = GFP_KERNEL;
+                page_mask = GFP_HIGHUSER;
        }
-        dma_addr = ib_dma_map_page(ic->i_cm_id->device,
+        if (!ic->i_cache_incs.ready)
-                                  ic->i_frag.f_page,
+                rds_ib_cache_xfer_to_ready(&ic->i_cache_incs);
-                                  ic->i_frag.f_offset,
+        if (!ic->i_cache_frags.ready)
-                                  RDS_FRAG_SIZE,
+                rds_ib_cache_xfer_to_ready(&ic->i_cache_frags);
-                                  DMA_FROM_DEVICE);
-        if (ib_dma_mapping_error(ic->i_cm_id->device, dma_addr))
-                goto out;
        /*
-         * Once we get the RDS_PAGE_LAST_OFF frag then rds_ib_frag_unmap()
+         * ibinc was taken from recv if recv contained the start of a message.
-         * must be called on this recv.  This happens as completions hit
+         * recvs that were continuations will still have this allocated.
-         * in order or on connection shutdown.
         */
-        recv->r_frag->f_page = ic->i_frag.f_page;
+        if (!recv->r_ibinc) {
-        recv->r_frag->f_offset = ic->i_frag.f_offset;
+                recv->r_ibinc = rds_ib_refill_one_inc(ic, slab_mask);
-        recv->r_frag->f_mapped = dma_addr;
+                if (!recv->r_ibinc)
+                        goto out;
+        }
-        sge = rds_ib_data_sge(ic, recv->r_sge);
+        WARN_ON(recv->r_frag); /* leak! */
-        sge->addr = dma_addr;
+        recv->r_frag = rds_ib_refill_one_frag(ic, slab_mask, page_mask);
-        sge->length = RDS_FRAG_SIZE;
+        if (!recv->r_frag)
+                goto out;
+        ret = ib_dma_map_sg(ic->i_cm_id->device, &recv->r_frag->f_sg,
+                            1, DMA_FROM_DEVICE);
+        WARN_ON(ret != 1);
-        sge = rds_ib_header_sge(ic, recv->r_sge);
+        sge = &recv->r_sge[0];
        sge->addr = ic->i_recv_hdrs_dma + (recv - ic->i_recvs) * sizeof(struct rds_header);
        sge->length = sizeof(struct rds_header);
-        get_page(recv->r_frag->f_page);
+        sge = &recv->r_sge[1];
+        sge->addr = sg_dma_address(&recv->r_frag->f_sg);
-        if (ic->i_frag.f_offset < RDS_PAGE_LAST_OFF) {
+        sge->length = sg_dma_len(&recv->r_frag->f_sg);
-                ic->i_frag.f_offset += RDS_FRAG_SIZE;
-        } else {
-                put_page(ic->i_frag.f_page);
-                ic->i_frag.f_page = NULL;
-                ic->i_frag.f_offset = 0;
-        }
        ret = 0;
 out:
@@ -216,13 +350,11 @@ out:
 /*
 * This tries to allocate and post unused work requests after making sure that
 * they have all the allocations they need to queue received fragments into
- * sockets.  The i_recv_mutex is held here so that ring_alloc and _unalloc
+ * sockets.
- * pairs don't go unmatched.
 *
 * -1 is returned if posting fails due to temporary resource exhaustion.
 */
-int rds_ib_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
+void rds_ib_recv_refill(struct rds_connection *conn, int prefill)
-                       gfp_t page_gfp, int prefill)
 {
        struct rds_ib_connection *ic = conn->c_transport_data;
        struct rds_ib_recv_work *recv;
@@ -236,28 +368,25 @@ int rds_ib_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
                if (pos >= ic->i_recv_ring.w_nr) {
                        printk(KERN_NOTICE "Argh - ring alloc returned pos=%u\n",
                                        pos);
-                        ret = -EINVAL;
                        break;
                }
                recv = &ic->i_recvs[pos];
-                ret = rds_ib_recv_refill_one(conn, recv, kptr_gfp, page_gfp);
+                ret = rds_ib_recv_refill_one(conn, recv, prefill);
                if (ret) {
-                        ret = -1;
                        break;
                }
                /* XXX when can this fail? */
                ret = ib_post_recv(ic->i_cm_id->qp, &recv->r_wr, &failed_wr);
                rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n", recv,
-                         recv->r_ibinc, recv->r_frag->f_page,
+                         recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
-                         (long) recv->r_frag->f_mapped, ret);
+                         (long) sg_dma_address(&recv->r_frag->f_sg), ret);
                if (ret) {
                        rds_ib_conn_error(conn, "recv post on "
                               "%pI4 returned %d, disconnecting and "
                               "reconnecting\n", &conn->c_faddr,
                               ret);
-                        ret = -1;
                        break;
                }
@@ -270,37 +399,73 @@ int rds_ib_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
        if (ret)
                rds_ib_ring_unalloc(&ic->i_recv_ring, 1);
-        return ret;
 }
-void rds_ib_inc_purge(struct rds_incoming *inc)
+/*
+ * We want to recycle several types of recv allocations, like incs and frags.
+ * To use this, the *_free() function passes in the ptr to a list_head within
+ * the recyclee, as well as the cache to put it on.
+ *
+ * First, we put the memory on a percpu list. When this reaches a certain size,
+ * We move it to an intermediate non-percpu list in a lockless manner, with some
+ * xchg/compxchg wizardry.
+ *
+ * N.B. Instead of a list_head as the anchor, we use a single pointer, which can
+ * be NULL and xchg'd. The list is actually empty when the pointer is NULL, and
+ * list_empty() will return true with one element is actually present.
+ */
+static void rds_ib_recv_cache_put(struct list_head *new_item,
+                                 struct rds_ib_refill_cache *cache)
 {
-        struct rds_ib_incoming *ibinc;
+        unsigned long flags;
-        struct rds_page_frag *frag;
+        struct rds_ib_cache_head *chp;
-        struct rds_page_frag *pos;
+        struct list_head *old;
-        ibinc = container_of(inc, struct rds_ib_incoming, ii_inc);
+        local_irq_save(flags);
-        rdsdebug("purging ibinc %p inc %p\n", ibinc, inc);
-        list_for_each_entry_safe(frag, pos, &ibinc->ii_frags, f_item) {
+        chp = per_cpu_ptr(cache->percpu, smp_processor_id());
-                list_del_init(&frag->f_item);
+        if (!chp->first)
-                rds_ib_frag_drop_page(frag);
+                INIT_LIST_HEAD(new_item);
-                rds_ib_frag_free(frag);
+        else /* put on front */
-        }
+                list_add_tail(new_item, chp->first);
+        chp->first = new_item;
+        chp->count++;
+        if (chp->count < RDS_IB_RECYCLE_BATCH_COUNT)
+                goto end;
+        /*
+         * Return our per-cpu first list to the cache's xfer by atomically
+         * grabbing the current xfer list, appending it to our per-cpu list,
+         * and then atomically returning that entire list back to the
+         * cache's xfer list as long as it's still empty.
+         */
+        do {
+                old = xchg(&cache->xfer, NULL);
+                if (old)
+                        list_splice_entire_tail(old, chp->first);
+                old = cmpxchg(&cache->xfer, NULL, chp->first);
+        } while (old);
+        chp->first = NULL;
+        chp->count = 0;
+end:
+        local_irq_restore(flags);
 }
-void rds_ib_inc_free(struct rds_incoming *inc)
+static struct list_head *rds_ib_recv_cache_get(struct rds_ib_refill_cache *cache)
 {
-        struct rds_ib_incoming *ibinc;
+        struct list_head *head = cache->ready;
-        ibinc = container_of(inc, struct rds_ib_incoming, ii_inc);
+        if (head) {
+                if (!list_empty(head)) {
+                        cache->ready = head->next;
+                        list_del_init(head);
+                } else
+                        cache->ready = NULL;
+        }
-        rds_ib_inc_purge(inc);
+        return head;
-        rdsdebug("freeing ibinc %p inc %p\n", ibinc, inc);
-        BUG_ON(!list_empty(&ibinc->ii_frags));
-        kmem_cache_free(rds_ib_incoming_slab, ibinc);
-        atomic_dec(&rds_ib_allocation);
-        BUG_ON(atomic_read(&rds_ib_allocation) < 0);
 }
 int rds_ib_inc_copy_to_user(struct rds_incoming *inc, struct iovec *first_iov,
@@ -336,13 +501,13 @@ int rds_ib_inc_copy_to_user(struct rds_incoming *inc, struct iovec *first_iov,
                to_copy = min_t(unsigned long, to_copy, len - copied);
                rdsdebug("%lu bytes to user [%p, %zu] + %lu from frag "
-                         "[%p, %lu] + %lu\n",
+                         "[%p, %u] + %lu\n",
                         to_copy, iov->iov_base, iov->iov_len, iov_off,
-                         frag->f_page, frag->f_offset, frag_off);
+                         sg_page(&frag->f_sg), frag->f_sg.offset, frag_off);
                /* XXX needs + offset for multiple recvs per page */
-                ret = rds_page_copy_to_user(frag->f_page,
+                ret = rds_page_copy_to_user(sg_page(&frag->f_sg),
-                                            frag->f_offset + frag_off,
+                                            frag->f_sg.offset + frag_off,
                                            iov->iov_base + iov_off,
                                            to_copy);
                if (ret) {
@@ -557,47 +722,6 @@ u64 rds_ib_piggyb_ack(struct rds_ib_connection *ic)
        return rds_ib_get_ack(ic);
 }
-static struct rds_header *rds_ib_get_header(struct rds_connection *conn,
-                                            struct rds_ib_recv_work *recv,
-                                            u32 data_len)
-{
-        struct rds_ib_connection *ic = conn->c_transport_data;
-        void *hdr_buff = &ic->i_recv_hdrs[recv - ic->i_recvs];
-        void *addr;
-        u32 misplaced_hdr_bytes;
-        /*
-         * Support header at the front (RDS 3.1+) as well as header-at-end.
-         *
-         * Cases:
-         * 1) header all in header buff (great!)
-         * 2) header all in data page (copy all to header buff)
-         * 3) header split across hdr buf + data page
-         *    (move bit in hdr buff to end before copying other bit from data page)
-         */
-        if (conn->c_version > RDS_PROTOCOL_3_0 || data_len == RDS_FRAG_SIZE)
-                return hdr_buff;
-        if (data_len <= (RDS_FRAG_SIZE - sizeof(struct rds_header))) {
-                addr = kmap_atomic(recv->r_frag->f_page, KM_SOFTIRQ0);
-                memcpy(hdr_buff,
-                       addr + recv->r_frag->f_offset + data_len,
-                       sizeof(struct rds_header));
-                kunmap_atomic(addr, KM_SOFTIRQ0);
-                return hdr_buff;
-        }
-        misplaced_hdr_bytes = (sizeof(struct rds_header) - (RDS_FRAG_SIZE - data_len));
-        memmove(hdr_buff + misplaced_hdr_bytes, hdr_buff, misplaced_hdr_bytes);
-        addr = kmap_atomic(recv->r_frag->f_page, KM_SOFTIRQ0);
-        memcpy(hdr_buff, addr + recv->r_frag->f_offset + data_len,
-               sizeof(struct rds_header) - misplaced_hdr_bytes);
-        kunmap_atomic(addr, KM_SOFTIRQ0);
-        return hdr_buff;
-}
 /*
 * It's kind of lame that we're copying from the posted receive pages into
 * long-lived bitmaps.  We could have posted the bitmaps and rdma written into
@@ -639,7 +763,7 @@ static void rds_ib_cong_recv(struct rds_connection *conn,
                to_copy = min(RDS_FRAG_SIZE - frag_off, PAGE_SIZE - map_off);
                BUG_ON(to_copy & 7); /* Must be 64bit aligned. */
-                addr = kmap_atomic(frag->f_page, KM_SOFTIRQ0);
+                addr = kmap_atomic(sg_page(&frag->f_sg), KM_SOFTIRQ0);
                src = addr + frag_off;
                dst = (void *)map->m_page_addrs[map_page] + map_off;
@@ -710,7 +834,7 @@ static void rds_ib_process_recv(struct rds_connection *conn,
        }
        data_len -= sizeof(struct rds_header);
-        ihdr = rds_ib_get_header(conn, recv, data_len);
+        ihdr = &ic->i_recv_hdrs[recv - ic->i_recvs];
        /* Validate the checksum. */
        if (!rds_message_verify_checksum(ihdr)) {
@@ -742,12 +866,12 @@ static void rds_ib_process_recv(struct rds_connection *conn,
                 * the inc is freed.  We don't go that route, so we have to drop the
                 * page ref ourselves.  We can't just leave the page on the recv
                 * because that confuses the dma mapping of pages and each recv's use
-                 * of a partial page.  We can leave the frag, though, it will be
+                 * of a partial page.
-                 * reused.
                 *
                 * FIXME: Fold this into the code path below.
                 */
-                rds_ib_frag_drop_page(recv->r_frag);
+                rds_ib_frag_free(ic, recv->r_frag);
+                recv->r_frag = NULL;
                return;
        }
@@ -757,7 +881,7 @@ static void rds_ib_process_recv(struct rds_connection *conn,
         * into the inc and save the inc so we can hang upcoming fragments
         * off its list.
         */
-        if (ibinc == NULL) {
+        if (!ibinc) {
                ibinc = recv->r_ibinc;
                recv->r_ibinc = NULL;
                ic->i_ibinc = ibinc;
@@ -842,32 +966,38 @@ static inline void rds_poll_cq(struct rds_ib_connection *ic,
        struct rds_ib_recv_work *recv;
        while (ib_poll_cq(ic->i_recv_cq, 1, &wc) > 0) {
-                rdsdebug("wc wr_id 0x%llx status %u byte_len %u imm_data %u\n",
+                rdsdebug("wc wr_id 0x%llx status %u (%s) byte_len %u imm_data %u\n",
-                         (unsigned long long)wc.wr_id, wc.status, wc.byte_len,
+                         (unsigned long long)wc.wr_id, wc.status,
+                         rds_ib_wc_status_str(wc.status), wc.byte_len,
                         be32_to_cpu(wc.ex.imm_data));
                rds_ib_stats_inc(s_ib_rx_cq_event);
                recv = &ic->i_recvs[rds_ib_ring_oldest(&ic->i_recv_ring)];
-                rds_ib_recv_unmap_page(ic, recv);
+                ib_dma_unmap_sg(ic->i_cm_id->device, &recv->r_frag->f_sg, 1, DMA_FROM_DEVICE);
                /*
                 * Also process recvs in connecting state because it is possible
                 * to get a recv completion _before_ the rdmacm ESTABLISHED
                 * event is processed.
                 */
-                if (rds_conn_up(conn) || rds_conn_connecting(conn)) {
+                if (wc.status == IB_WC_SUCCESS) {
+                        rds_ib_process_recv(conn, recv, wc.byte_len, state);
+                } else {
                        /* We expect errors as the qp is drained during shutdown */
-                        if (wc.status == IB_WC_SUCCESS) {
+                        if (rds_conn_up(conn) || rds_conn_connecting(conn))
-                                rds_ib_process_recv(conn, recv, wc.byte_len, state);
+                                rds_ib_conn_error(conn, "recv completion on %pI4 had "
-                        } else {
+                                                  "status %u (%s), disconnecting and "
-                                rds_ib_conn_error(conn, "recv completion on "
+                                                  "reconnecting\n", &conn->c_faddr,
-                                       "%pI4 had status %u, disconnecting and "
+                                                  wc.status,
-                                       "reconnecting\n", &conn->c_faddr,
+                                                  rds_ib_wc_status_str(wc.status));
-                                       wc.status);
-                        }
                }
+                /*
+                 * It's very important that we only free this ring entry if we've truly
+                 * freed the resources allocated to the entry.  The refilling path can
+                 * leak if we don't.
+                 */
                rds_ib_ring_free(&ic->i_recv_ring, 1);
        }
 }
@@ -897,11 +1027,8 @@ void rds_ib_recv_tasklet_fn(unsigned long data)
        if (rds_ib_ring_empty(&ic->i_recv_ring))
                rds_ib_stats_inc(s_ib_rx_ring_empty);
-        /*
-         * If the ring is running low, then schedule the thread to refill.
-         */
        if (rds_ib_ring_low(&ic->i_recv_ring))
-                queue_delayed_work(rds_wq, &conn->c_recv_w, 0);
+                rds_ib_recv_refill(conn, 0);
 }
 int rds_ib_recv(struct rds_connection *conn)
@@ -910,25 +1037,13 @@ int rds_ib_recv(struct rds_connection *conn)
        int ret = 0;
        rdsdebug("conn %p\n", conn);
-        /*
-         * If we get a temporary posting failure in this context then
-         * we're really low and we want the caller to back off for a bit.
-         */
-        mutex_lock(&ic->i_recv_mutex);
-        if (rds_ib_recv_refill(conn, GFP_KERNEL, GFP_HIGHUSER, 0))
-                ret = -ENOMEM;
-        else
-                rds_ib_stats_inc(s_ib_rx_refill_from_thread);
-        mutex_unlock(&ic->i_recv_mutex);
        if (rds_conn_up(conn))
                rds_ib_attempt_ack(ic);
        return ret;
 }
-int __init rds_ib_recv_init(void)
+int rds_ib_recv_init(void)
 {
        struct sysinfo si;
        int ret = -ENOMEM;
@@ -939,14 +1054,14 @@ int __init rds_ib_recv_init(void)
        rds_ib_incoming_slab = kmem_cache_create("rds_ib_incoming",
                                        sizeof(struct rds_ib_incoming),
-                                        0, 0, NULL);
+                                        0, SLAB_HWCACHE_ALIGN, NULL);
-        if (rds_ib_incoming_slab == NULL)
+        if (!rds_ib_incoming_slab)
                goto out;
        rds_ib_frag_slab = kmem_cache_create("rds_ib_frag",
                                        sizeof(struct rds_page_frag),
-                                        0, 0, NULL);
+                                        0, SLAB_HWCACHE_ALIGN, NULL);
-        if (rds_ib_frag_slab == NULL)
+        if (!rds_ib_frag_slab)
                kmem_cache_destroy(rds_ib_incoming_slab);
        else
                ret = 0;
diff --git a/net/rds/ib_send.c b/net/rds/ib_send.c
index 17fa80803ab0..71f373c421bc 100644
--- a/net/rds/ib_send.c
+++ b/net/rds/ib_send.c
@@ -36,11 +36,49 @@
 #include <linux/dmapool.h>
 #include "rds.h"
-#include "rdma.h"
 #include "ib.h"
-static void rds_ib_send_rdma_complete(struct rds_message *rm,
+static char *rds_ib_wc_status_strings[] = {
-                                      int wc_status)
+#define RDS_IB_WC_STATUS_STR(foo) \
+                [IB_WC_##foo] = __stringify(IB_WC_##foo)
+        RDS_IB_WC_STATUS_STR(SUCCESS),
+        RDS_IB_WC_STATUS_STR(LOC_LEN_ERR),
+        RDS_IB_WC_STATUS_STR(LOC_QP_OP_ERR),
+        RDS_IB_WC_STATUS_STR(LOC_EEC_OP_ERR),
+        RDS_IB_WC_STATUS_STR(LOC_PROT_ERR),
+        RDS_IB_WC_STATUS_STR(WR_FLUSH_ERR),
+        RDS_IB_WC_STATUS_STR(MW_BIND_ERR),
+        RDS_IB_WC_STATUS_STR(BAD_RESP_ERR),
+        RDS_IB_WC_STATUS_STR(LOC_ACCESS_ERR),
+        RDS_IB_WC_STATUS_STR(REM_INV_REQ_ERR),
+        RDS_IB_WC_STATUS_STR(REM_ACCESS_ERR),
+        RDS_IB_WC_STATUS_STR(REM_OP_ERR),
+        RDS_IB_WC_STATUS_STR(RETRY_EXC_ERR),
+        RDS_IB_WC_STATUS_STR(RNR_RETRY_EXC_ERR),
+        RDS_IB_WC_STATUS_STR(LOC_RDD_VIOL_ERR),
+        RDS_IB_WC_STATUS_STR(REM_INV_RD_REQ_ERR),
+        RDS_IB_WC_STATUS_STR(REM_ABORT_ERR),
+        RDS_IB_WC_STATUS_STR(INV_EECN_ERR),
+        RDS_IB_WC_STATUS_STR(INV_EEC_STATE_ERR),
+        RDS_IB_WC_STATUS_STR(FATAL_ERR),
+        RDS_IB_WC_STATUS_STR(RESP_TIMEOUT_ERR),
+        RDS_IB_WC_STATUS_STR(GENERAL_ERR),
+#undef RDS_IB_WC_STATUS_STR
+};
+char *rds_ib_wc_status_str(enum ib_wc_status status)
+{
+        return rds_str_array(rds_ib_wc_status_strings,
+                             ARRAY_SIZE(rds_ib_wc_status_strings), status);
+}
+/*
+ * Convert IB-specific error message to RDS error message and call core
+ * completion handler.
+ */
+static void rds_ib_send_complete(struct rds_message *rm,
+                                 int wc_status,
+                                 void (*complete)(struct rds_message *rm, int status))
 {
        int notify_status;
@@ -60,69 +98,125 @@ static void rds_ib_send_rdma_complete(struct rds_message *rm,
                notify_status = RDS_RDMA_OTHER_ERROR;
                break;
        }
-        rds_rdma_send_complete(rm, notify_status);
+        complete(rm, notify_status);
+}
+static void rds_ib_send_unmap_data(struct rds_ib_connection *ic,
+                                   struct rm_data_op *op,
+                                   int wc_status)
+{
+        if (op->op_nents)
+                ib_dma_unmap_sg(ic->i_cm_id->device,
+                                op->op_sg, op->op_nents,
+                                DMA_TO_DEVICE);
 }
 static void rds_ib_send_unmap_rdma(struct rds_ib_connection *ic,
-                                   struct rds_rdma_op *op)
+                                   struct rm_rdma_op *op,
+                                   int wc_status)
 {
-        if (op->r_mapped) {
+        if (op->op_mapped) {
                ib_dma_unmap_sg(ic->i_cm_id->device,
-                        op->r_sg, op->r_nents,
+                                op->op_sg, op->op_nents,
-                        op->r_write ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
+                                op->op_write ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
-                op->r_mapped = 0;
+                op->op_mapped = 0;
        }
+        /* If the user asked for a completion notification on this
+         * message, we can implement three different semantics:
+         *  1.  Notify when we received the ACK on the RDS message
+         *      that was queued with the RDMA. This provides reliable
+         *      notification of RDMA status at the expense of a one-way
+         *      packet delay.
+         *  2.  Notify when the IB stack gives us the completion event for
+         *      the RDMA operation.
+         *  3.  Notify when the IB stack gives us the completion event for
+         *      the accompanying RDS messages.
+         * Here, we implement approach #3. To implement approach #2,
+         * we would need to take an event for the rdma WR. To implement #1,
+         * don't call rds_rdma_send_complete at all, and fall back to the notify
+         * handling in the ACK processing code.
+         *
+         * Note: There's no need to explicitly sync any RDMA buffers using
+         * ib_dma_sync_sg_for_cpu - the completion for the RDMA
+         * operation itself unmapped the RDMA buffers, which takes care
+         * of synching.
+         */
+        rds_ib_send_complete(container_of(op, struct rds_message, rdma),
+                             wc_status, rds_rdma_send_complete);
+        if (op->op_write)
+                rds_stats_add(s_send_rdma_bytes, op->op_bytes);
+        else
+                rds_stats_add(s_recv_rdma_bytes, op->op_bytes);
 }
-static void rds_ib_send_unmap_rm(struct rds_ib_connection *ic,
+static void rds_ib_send_unmap_atomic(struct rds_ib_connection *ic,
-                          struct rds_ib_send_work *send,
+                                     struct rm_atomic_op *op,
-                          int wc_status)
+                                     int wc_status)
 {
-        struct rds_message *rm = send->s_rm;
+        /* unmap atomic recvbuf */
+        if (op->op_mapped) {
-        rdsdebug("ic %p send %p rm %p\n", ic, send, rm);
+                ib_dma_unmap_sg(ic->i_cm_id->device, op->op_sg, 1,
+                                DMA_FROM_DEVICE);
-        ib_dma_unmap_sg(ic->i_cm_id->device,
+                op->op_mapped = 0;
-                     rm->m_sg, rm->m_nents,
+        }
-                     DMA_TO_DEVICE);
-        if (rm->m_rdma_op != NULL) {
-                rds_ib_send_unmap_rdma(ic, rm->m_rdma_op);
-                /* If the user asked for a completion notification on this
-                 * message, we can implement three different semantics:
-                 *  1.  Notify when we received the ACK on the RDS message
-                 *      that was queued with the RDMA. This provides reliable
-                 *      notification of RDMA status at the expense of a one-way
-                 *      packet delay.
-                 *  2.  Notify when the IB stack gives us the completion event for
-                 *      the RDMA operation.
-                 *  3.  Notify when the IB stack gives us the completion event for
-                 *      the accompanying RDS messages.
-                 * Here, we implement approach #3. To implement approach #2,
-                 * call rds_rdma_send_complete from the cq_handler. To implement #1,
-                 * don't call rds_rdma_send_complete at all, and fall back to the notify
-                 * handling in the ACK processing code.
-                 *
-                 * Note: There's no need to explicitly sync any RDMA buffers using
-                 * ib_dma_sync_sg_for_cpu - the completion for the RDMA
-                 * operation itself unmapped the RDMA buffers, which takes care
-                 * of synching.
-                 */
-                rds_ib_send_rdma_complete(rm, wc_status);
-                if (rm->m_rdma_op->r_write)
+        rds_ib_send_complete(container_of(op, struct rds_message, atomic),
-                        rds_stats_add(s_send_rdma_bytes, rm->m_rdma_op->r_bytes);
+                             wc_status, rds_atomic_send_complete);
-                else
-                        rds_stats_add(s_recv_rdma_bytes, rm->m_rdma_op->r_bytes);
+        if (op->op_type == RDS_ATOMIC_TYPE_CSWP)
+                rds_ib_stats_inc(s_ib_atomic_cswp);
+        else
+                rds_ib_stats_inc(s_ib_atomic_fadd);
+}
+/*
+ * Unmap the resources associated with a struct send_work.
+ *
+ * Returns the rm for no good reason other than it is unobtainable
+ * other than by switching on wr.opcode, currently, and the caller,
+ * the event handler, needs it.
+ */
+static struct rds_message *rds_ib_send_unmap_op(struct rds_ib_connection *ic,
+                                                struct rds_ib_send_work *send,
+                                                int wc_status)
+{
+        struct rds_message *rm = NULL;
+        /* In the error case, wc.opcode sometimes contains garbage */
+        switch (send->s_wr.opcode) {
+        case IB_WR_SEND:
+                if (send->s_op) {
+                        rm = container_of(send->s_op, struct rds_message, data);
+                        rds_ib_send_unmap_data(ic, send->s_op, wc_status);
+                }
+                break;
+        case IB_WR_RDMA_WRITE:
+        case IB_WR_RDMA_READ:
+                if (send->s_op) {
+                        rm = container_of(send->s_op, struct rds_message, rdma);
+                        rds_ib_send_unmap_rdma(ic, send->s_op, wc_status);
+                }
+                break;
+        case IB_WR_ATOMIC_FETCH_AND_ADD:
+        case IB_WR_ATOMIC_CMP_AND_SWP:
+                if (send->s_op) {
+                        rm = container_of(send->s_op, struct rds_message, atomic);
+                        rds_ib_send_unmap_atomic(ic, send->s_op, wc_status);
+                }
+                break;
+        default:
+                if (printk_ratelimit())
+                        printk(KERN_NOTICE
+                               "RDS/IB: %s: unexpected opcode 0x%x in WR!\n",
+                               __func__, send->s_wr.opcode);
+                break;
        }
-        /* If anyone waited for this message to get flushed out, wake
+        send->s_wr.opcode = 0xdead;
-         * them up now */
-        rds_message_unmapped(rm);
-        rds_message_put(rm);
+        return rm;
-        send->s_rm = NULL;
 }
 void rds_ib_send_init_ring(struct rds_ib_connection *ic)
@@ -133,23 +227,18 @@ void rds_ib_send_init_ring(struct rds_ib_connection *ic)
        for (i = 0, send = ic->i_sends; i < ic->i_send_ring.w_nr; i++, send++) {
                struct ib_sge *sge;
-                send->s_rm = NULL;
                send->s_op = NULL;
                send->s_wr.wr_id = i;
                send->s_wr.sg_list = send->s_sge;
-                send->s_wr.num_sge = 1;
-                send->s_wr.opcode = IB_WR_SEND;
-                send->s_wr.send_flags = 0;
                send->s_wr.ex.imm_data = 0;
-                sge = rds_ib_data_sge(ic, send->s_sge);
+                sge = &send->s_sge[0];
-                sge->lkey = ic->i_mr->lkey;
-                sge = rds_ib_header_sge(ic, send->s_sge);
                sge->addr = ic->i_send_hdrs_dma + (i * sizeof(struct rds_header));
                sge->length = sizeof(struct rds_header);
                sge->lkey = ic->i_mr->lkey;
+                send->s_sge[1].lkey = ic->i_mr->lkey;
        }
 }
@@ -159,16 +248,24 @@ void rds_ib_send_clear_ring(struct rds_ib_connection *ic)
        u32 i;
        for (i = 0, send = ic->i_sends; i < ic->i_send_ring.w_nr; i++, send++) {
-                if (send->s_wr.opcode == 0xdead)
+                if (send->s_op && send->s_wr.opcode != 0xdead)
-                        continue;
+                        rds_ib_send_unmap_op(ic, send, IB_WC_WR_FLUSH_ERR);
-                if (send->s_rm)
-                        rds_ib_send_unmap_rm(ic, send, IB_WC_WR_FLUSH_ERR);
-                if (send->s_op)
-                        rds_ib_send_unmap_rdma(ic, send->s_op);
        }
 }
 /*
+ * The only fast path caller always has a non-zero nr, so we don't
+ * bother testing nr before performing the atomic sub.
+ */
+static void rds_ib_sub_signaled(struct rds_ib_connection *ic, int nr)
+{
+        if ((atomic_sub_return(nr, &ic->i_signaled_sends) == 0) &&
+            waitqueue_active(&rds_ib_ring_empty_wait))
+                wake_up(&rds_ib_ring_empty_wait);
+        BUG_ON(atomic_read(&ic->i_signaled_sends) < 0);
+}
+/*
 * The _oldest/_free ring operations here race cleanly with the alloc/unalloc
 * operations performed in the send path.  As the sender allocs and potentially
 * unallocs the next free entry in the ring it doesn't alter which is
@@ -178,12 +275,14 @@ void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context)
 {
        struct rds_connection *conn = context;
        struct rds_ib_connection *ic = conn->c_transport_data;
+        struct rds_message *rm = NULL;
        struct ib_wc wc;
        struct rds_ib_send_work *send;
        u32 completed;
        u32 oldest;
        u32 i = 0;
        int ret;
+        int nr_sig = 0;
        rdsdebug("cq %p conn %p\n", cq, conn);
        rds_ib_stats_inc(s_ib_tx_cq_call);
@@ -192,8 +291,9 @@ void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context)
                rdsdebug("ib_req_notify_cq send failed: %d\n", ret);
        while (ib_poll_cq(cq, 1, &wc) > 0) {
-                rdsdebug("wc wr_id 0x%llx status %u byte_len %u imm_data %u\n",
+                rdsdebug("wc wr_id 0x%llx status %u (%s) byte_len %u imm_data %u\n",
-                         (unsigned long long)wc.wr_id, wc.status, wc.byte_len,
+                         (unsigned long long)wc.wr_id, wc.status,
+                         rds_ib_wc_status_str(wc.status), wc.byte_len,
                         be32_to_cpu(wc.ex.imm_data));
                rds_ib_stats_inc(s_ib_tx_cq_event);
@@ -210,51 +310,30 @@ void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context)
                for (i = 0; i < completed; i++) {
                        send = &ic->i_sends[oldest];
+                        if (send->s_wr.send_flags & IB_SEND_SIGNALED)
+                                nr_sig++;
-                        /* In the error case, wc.opcode sometimes contains garbage */
+                        rm = rds_ib_send_unmap_op(ic, send, wc.status);
-                        switch (send->s_wr.opcode) {
-                        case IB_WR_SEND:
-                                if (send->s_rm)
-                                        rds_ib_send_unmap_rm(ic, send, wc.status);
-                                break;
-                        case IB_WR_RDMA_WRITE:
-                        case IB_WR_RDMA_READ:
-                                /* Nothing to be done - the SG list will be unmapped
-                                 * when the SEND completes. */
-                                break;
-                        default:
-                                if (printk_ratelimit())
-                                        printk(KERN_NOTICE
-                                                "RDS/IB: %s: unexpected opcode 0x%x in WR!\n",
-                                                __func__, send->s_wr.opcode);
-                                break;
-                        }
-                        send->s_wr.opcode = 0xdead;
-                        send->s_wr.num_sge = 1;
                        if (send->s_queued + HZ/2 < jiffies)
                                rds_ib_stats_inc(s_ib_tx_stalled);
-                        /* If a RDMA operation produced an error, signal this right
+                        if (send->s_op) {
-                         * away. If we don't, the subsequent SEND that goes with this
+                                if (send->s_op == rm->m_final_op) {
-                         * RDMA will be canceled with ERR_WFLUSH, and the application
+                                        /* If anyone waited for this message to get flushed out, wake
-                         * never learn that the RDMA failed. */
+                                         * them up now */
-                        if (unlikely(wc.status == IB_WC_REM_ACCESS_ERR && send->s_op)) {
+                                        rds_message_unmapped(rm);
-                                struct rds_message *rm;
-                                rm = rds_send_get_message(conn, send->s_op);
-                                if (rm) {
-                                        if (rm->m_rdma_op)
-                                                rds_ib_send_unmap_rdma(ic, rm->m_rdma_op);
-                                        rds_ib_send_rdma_complete(rm, wc.status);
-                                        rds_message_put(rm);
                                }
+                                rds_message_put(rm);
+                                send->s_op = NULL;
                        }
                        oldest = (oldest + 1) % ic->i_send_ring.w_nr;
                }
                rds_ib_ring_free(&ic->i_send_ring, completed);
+                rds_ib_sub_signaled(ic, nr_sig);
+                nr_sig = 0;
                if (test_and_clear_bit(RDS_LL_SEND_FULL, &conn->c_flags) ||
                    test_bit(0, &conn->c_map_queued))
@@ -262,10 +341,10 @@ void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context)
                /* We expect errors as the qp is drained during shutdown */
                if (wc.status != IB_WC_SUCCESS && rds_conn_up(conn)) {
-                        rds_ib_conn_error(conn,
+                        rds_ib_conn_error(conn, "send completion on %pI4 had status "
-                                "send completion on %pI4 "
+                                          "%u (%s), disconnecting and reconnecting\n",
-                                "had status %u, disconnecting and reconnecting\n",
+                                          &conn->c_faddr, wc.status,
-                                &conn->c_faddr, wc.status);
+                                          rds_ib_wc_status_str(wc.status));
                }
        }
 }
@@ -294,7 +373,7 @@ void rds_ib_send_cq_comp_handler(struct ib_cq *cq, void *context)
 * credits (see rds_ib_send_add_credits below).
 *
 * The RDS send code is essentially single-threaded; rds_send_xmit
- * grabs c_send_lock to ensure exclusive access to the send ring.
+ * sets RDS_IN_XMIT to ensure exclusive access to the send ring.
 * However, the ACK sending code is independent and can race with
 * message SENDs.
 *
@@ -413,40 +492,21 @@ void rds_ib_advertise_credits(struct rds_connection *conn, unsigned int posted)
                set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
 }
-static inline void
+static inline int rds_ib_set_wr_signal_state(struct rds_ib_connection *ic,
-rds_ib_xmit_populate_wr(struct rds_ib_connection *ic,
+                                             struct rds_ib_send_work *send,
-                struct rds_ib_send_work *send, unsigned int pos,
+                                             bool notify)
-                unsigned long buffer, unsigned int length,
-                int send_flags)
 {
-        struct ib_sge *sge;
+        /*
+         * We want to delay signaling completions just enough to get
-        WARN_ON(pos != send - ic->i_sends);
+         * the batching benefits but not so much that we create dead time
+         * on the wire.
-        send->s_wr.send_flags = send_flags;
+         */
-        send->s_wr.opcode = IB_WR_SEND;
+        if (ic->i_unsignaled_wrs-- == 0 || notify) {
-        send->s_wr.num_sge = 2;
+                ic->i_unsignaled_wrs = rds_ib_sysctl_max_unsig_wrs;
-        send->s_wr.next = NULL;
+                send->s_wr.send_flags |= IB_SEND_SIGNALED;
-        send->s_queued = jiffies;
+                return 1;
-        send->s_op = NULL;
-        if (length != 0) {
-                sge = rds_ib_data_sge(ic, send->s_sge);
-                sge->addr = buffer;
-                sge->length = length;
-                sge->lkey = ic->i_mr->lkey;
-                sge = rds_ib_header_sge(ic, send->s_sge);
-        } else {
-                /* We're sending a packet with no payload. There is only
-                 * one SGE */
-                send->s_wr.num_sge = 1;
-                sge = &send->s_sge[0];
        }
+        return 0;
-        sge->addr = ic->i_send_hdrs_dma + (pos * sizeof(struct rds_header));
-        sge->length = sizeof(struct rds_header);
-        sge->lkey = ic->i_mr->lkey;
 }
 /*
@@ -475,13 +535,14 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
        u32 pos;
        u32 i;
        u32 work_alloc;
-        u32 credit_alloc;
+        u32 credit_alloc = 0;
        u32 posted;
        u32 adv_credits = 0;
        int send_flags = 0;
-        int sent;
+        int bytes_sent = 0;
        int ret;
        int flow_controlled = 0;
+        int nr_sig = 0;
        BUG_ON(off % RDS_FRAG_SIZE);
        BUG_ON(hdr_off != 0 && hdr_off != sizeof(struct rds_header));
@@ -507,14 +568,13 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
                goto out;
        }
-        credit_alloc = work_alloc;
        if (ic->i_flowctl) {
                credit_alloc = rds_ib_send_grab_credits(ic, work_alloc, &posted, 0, RDS_MAX_ADV_CREDIT);
                adv_credits += posted;
                if (credit_alloc < work_alloc) {
                        rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc - credit_alloc);
                        work_alloc = credit_alloc;
-                        flow_controlled++;
+                        flow_controlled = 1;
                }
                if (work_alloc == 0) {
                        set_bit(RDS_LL_SEND_FULL, &conn->c_flags);
@@ -525,31 +585,25 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
        }
        /* map the message the first time we see it */
-        if (ic->i_rm == NULL) {
+        if (!ic->i_data_op) {
-                /*
+                if (rm->data.op_nents) {
-                printk(KERN_NOTICE "rds_ib_xmit prep msg dport=%u flags=0x%x len=%d\n",
+                        rm->data.op_count = ib_dma_map_sg(dev,
-                                be16_to_cpu(rm->m_inc.i_hdr.h_dport),
+                                                          rm->data.op_sg,
-                                rm->m_inc.i_hdr.h_flags,
+                                                          rm->data.op_nents,
-                                be32_to_cpu(rm->m_inc.i_hdr.h_len));
+                                                          DMA_TO_DEVICE);
-                   */
+                        rdsdebug("ic %p mapping rm %p: %d\n", ic, rm, rm->data.op_count);
-                if (rm->m_nents) {
+                        if (rm->data.op_count == 0) {
-                        rm->m_count = ib_dma_map_sg(dev,
-                                         rm->m_sg, rm->m_nents, DMA_TO_DEVICE);
-                        rdsdebug("ic %p mapping rm %p: %d\n", ic, rm, rm->m_count);
-                        if (rm->m_count == 0) {
                                rds_ib_stats_inc(s_ib_tx_sg_mapping_failure);
                                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
                                ret = -ENOMEM; /* XXX ? */
                                goto out;
                        }
                } else {
-                        rm->m_count = 0;
+                        rm->data.op_count = 0;
                }
-                ic->i_unsignaled_wrs = rds_ib_sysctl_max_unsig_wrs;
-                ic->i_unsignaled_bytes = rds_ib_sysctl_max_unsig_bytes;
                rds_message_addref(rm);
-                ic->i_rm = rm;
+                ic->i_data_op = &rm->data;
                /* Finalize the header */
                if (test_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags))
@@ -559,10 +613,10 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
                /* If it has a RDMA op, tell the peer we did it. This is
                 * used by the peer to release use-once RDMA MRs. */
-                if (rm->m_rdma_op) {
+                if (rm->rdma.op_active) {
                        struct rds_ext_header_rdma ext_hdr;
-                        ext_hdr.h_rdma_rkey = cpu_to_be32(rm->m_rdma_op->r_key);
+                        ext_hdr.h_rdma_rkey = cpu_to_be32(rm->rdma.op_rkey);
                        rds_message_add_extension(&rm->m_inc.i_hdr,
                                        RDS_EXTHDR_RDMA, &ext_hdr, sizeof(ext_hdr));
                }
@@ -582,99 +636,77 @@ int rds_ib_xmit(struct rds_connection *conn, struct rds_message *rm,
                /*
                 * Update adv_credits since we reset the ACK_REQUIRED bit.
                 */
-                rds_ib_send_grab_credits(ic, 0, &posted, 1, RDS_MAX_ADV_CREDIT - adv_credits);
+                if (ic->i_flowctl) {
-                adv_credits += posted;
+                        rds_ib_send_grab_credits(ic, 0, &posted, 1, RDS_MAX_ADV_CREDIT - adv_credits);
-                BUG_ON(adv_credits > 255);
+                        adv_credits += posted;
+                        BUG_ON(adv_credits > 255);
+                }
        }
-        send = &ic->i_sends[pos];
-        first = send;
-        prev = NULL;
-        scat = &rm->m_sg[sg];
-        sent = 0;
-        i = 0;
        /* Sometimes you want to put a fence between an RDMA
         * READ and the following SEND.
         * We could either do this all the time
         * or when requested by the user. Right now, we let
         * the application choose.
         */
-        if (rm->m_rdma_op && rm->m_rdma_op->r_fence)
+        if (rm->rdma.op_active && rm->rdma.op_fence)
                send_flags = IB_SEND_FENCE;
-        /*
+        /* Each frag gets a header. Msgs may be 0 bytes */
-         * We could be copying the header into the unused tail of the page.
+        send = &ic->i_sends[pos];
-         * That would need to be changed in the future when those pages might
+        first = send;
-         * be mapped userspace pages or page cache pages.  So instead we always
+        prev = NULL;
-         * use a second sge and our long-lived ring of mapped headers.  We send
+        scat = &ic->i_data_op->op_sg[sg];
-         * the header after the data so that the data payload can be aligned on
+        i = 0;
-         * the receiver.
+        do {
-         */
+                unsigned int len = 0;
-        /* handle a 0-len message */
+                /* Set up the header */
-        if (be32_to_cpu(rm->m_inc.i_hdr.h_len) == 0) {
+                send->s_wr.send_flags = send_flags;
-                rds_ib_xmit_populate_wr(ic, send, pos, 0, 0, send_flags);
+                send->s_wr.opcode = IB_WR_SEND;
-                goto add_header;
+                send->s_wr.num_sge = 1;
-        }
+                send->s_wr.next = NULL;
+                send->s_queued = jiffies;
+                send->s_op = NULL;
-        /* if there's data reference it with a chain of work reqs */
+                send->s_sge[0].addr = ic->i_send_hdrs_dma
-        for (; i < work_alloc && scat != &rm->m_sg[rm->m_count]; i++) {
+                        + (pos * sizeof(struct rds_header));
-                unsigned int len;
+                send->s_sge[0].length = sizeof(struct rds_header);
-                send = &ic->i_sends[pos];
+                memcpy(&ic->i_send_hdrs[pos], &rm->m_inc.i_hdr, sizeof(struct rds_header));
-                len = min(RDS_FRAG_SIZE, ib_sg_dma_len(dev, scat) - off);
+                /* Set up the data, if present */
-                rds_ib_xmit_populate_wr(ic, send, pos,
+                if (i < work_alloc
-                                ib_sg_dma_address(dev, scat) + off, len,
+                    && scat != &rm->data.op_sg[rm->data.op_count]) {
-                                send_flags);
+                        len = min(RDS_FRAG_SIZE, ib_sg_dma_len(dev, scat) - off);
+                        send->s_wr.num_sge = 2;
-                /*
+                        send->s_sge[1].addr = ib_sg_dma_address(dev, scat) + off;
-                 * We want to delay signaling completions just enough to get
+                        send->s_sge[1].length = len;
-                 * the batching benefits but not so much that we create dead time
-                 * on the wire.
-                 */
-                if (ic->i_unsignaled_wrs-- == 0) {
-                        ic->i_unsignaled_wrs = rds_ib_sysctl_max_unsig_wrs;
-                        send->s_wr.send_flags |= IB_SEND_SIGNALED | IB_SEND_SOLICITED;
-                }
-                ic->i_unsignaled_bytes -= len;
+                        bytes_sent += len;
-                if (ic->i_unsignaled_bytes <= 0) {
+                        off += len;
-                        ic->i_unsignaled_bytes = rds_ib_sysctl_max_unsig_bytes;
+                        if (off == ib_sg_dma_len(dev, scat)) {
-                        send->s_wr.send_flags |= IB_SEND_SIGNALED | IB_SEND_SOLICITED;
+                                scat++;
+                                off = 0;
+                        }
                }
+                rds_ib_set_wr_signal_state(ic, send, 0);
                /*
                 * Always signal the last one if we're stopping due to flow control.
                 */
-                if (flow_controlled && i == (work_alloc-1))
+                if (ic->i_flowctl && flow_controlled && i == (work_alloc-1))
                        send->s_wr.send_flags |= IB_SEND_SIGNALED | IB_SEND_SOLICITED;
+                if (send->s_wr.send_flags & IB_SEND_SIGNALED)
+                        nr_sig++;
                rdsdebug("send %p wr %p num_sge %u next %p\n", send,
                         &send->s_wr, send->s_wr.num_sge, send->s_wr.next);
-                sent += len;
+                if (ic->i_flowctl && adv_credits) {
-                off += len;
-                if (off == ib_sg_dma_len(dev, scat)) {
-                        scat++;
-                        off = 0;
-                }
-add_header:
-                /* Tack on the header after the data. The header SGE should already
-                 * have been set up to point to the right header buffer. */
-                memcpy(&ic->i_send_hdrs[pos], &rm->m_inc.i_hdr, sizeof(struct rds_header));
-                if (0) {
-                        struct rds_header *hdr = &ic->i_send_hdrs[pos];
-                        printk(KERN_NOTICE "send WR dport=%u flags=0x%x len=%d\n",
-                                be16_to_cpu(hdr->h_dport),
-                                hdr->h_flags,
-                                be32_to_cpu(hdr->h_len));
-                }
-                if (adv_credits) {
                        struct rds_header *hdr = &ic->i_send_hdrs[pos];
                        /* add credit and redo the header checksum */
@@ -689,20 +721,25 @@ add_header:
                prev = send;
                pos = (pos + 1) % ic->i_send_ring.w_nr;
-        }
+                send = &ic->i_sends[pos];
+                i++;
+        } while (i < work_alloc
+                 && scat != &rm->data.op_sg[rm->data.op_count]);
        /* Account the RDS header in the number of bytes we sent, but just once.
         * The caller has no concept of fragmentation. */
        if (hdr_off == 0)
-                sent += sizeof(struct rds_header);
+                bytes_sent += sizeof(struct rds_header);
        /* if we finished the message then send completion owns it */
-        if (scat == &rm->m_sg[rm->m_count]) {
+        if (scat == &rm->data.op_sg[rm->data.op_count]) {
-                prev->s_rm = ic->i_rm;
+                prev->s_op = ic->i_data_op;
-                prev->s_wr.send_flags |= IB_SEND_SIGNALED | IB_SEND_SOLICITED;
+                prev->s_wr.send_flags |= IB_SEND_SOLICITED;
-                ic->i_rm = NULL;
+                ic->i_data_op = NULL;
        }
+        /* Put back wrs & credits we didn't use */
        if (i < work_alloc) {
                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc - i);
                work_alloc = i;
@@ -710,6 +747,9 @@ add_header:
        if (ic->i_flowctl && i < credit_alloc)
                rds_ib_send_add_credits(conn, credit_alloc - i);
+        if (nr_sig)
+                atomic_add(nr_sig, &ic->i_signaled_sends);
        /* XXX need to worry about failed_wr and partial sends. */
        failed_wr = &first->s_wr;
        ret = ib_post_send(ic->i_cm_id->qp, &first->s_wr, &failed_wr);
@@ -720,32 +760,127 @@ add_header:
                printk(KERN_WARNING "RDS/IB: ib_post_send to %pI4 "
                       "returned %d\n", &conn->c_faddr, ret);
                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
-                if (prev->s_rm) {
+                rds_ib_sub_signaled(ic, nr_sig);
-                        ic->i_rm = prev->s_rm;
+                if (prev->s_op) {
-                        prev->s_rm = NULL;
+                        ic->i_data_op = prev->s_op;
+                        prev->s_op = NULL;
                }
                rds_ib_conn_error(ic->conn, "ib_post_send failed\n");
                goto out;
        }
-        ret = sent;
+        ret = bytes_sent;
 out:
        BUG_ON(adv_credits);
        return ret;
 }
-int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
+/*
+ * Issue atomic operation.
+ * A simplified version of the rdma case, we always map 1 SG, and
+ * only 8 bytes, for the return value from the atomic operation.
+ */
+int rds_ib_xmit_atomic(struct rds_connection *conn, struct rm_atomic_op *op)
+{
+        struct rds_ib_connection *ic = conn->c_transport_data;
+        struct rds_ib_send_work *send = NULL;
+        struct ib_send_wr *failed_wr;
+        struct rds_ib_device *rds_ibdev;
+        u32 pos;
+        u32 work_alloc;
+        int ret;
+        int nr_sig = 0;
+        rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
+        work_alloc = rds_ib_ring_alloc(&ic->i_send_ring, 1, &pos);
+        if (work_alloc != 1) {
+                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
+                rds_ib_stats_inc(s_ib_tx_ring_full);
+                ret = -ENOMEM;
+                goto out;
+        }
+        /* address of send request in ring */
+        send = &ic->i_sends[pos];
+        send->s_queued = jiffies;
+        if (op->op_type == RDS_ATOMIC_TYPE_CSWP) {
+                send->s_wr.opcode = IB_WR_MASKED_ATOMIC_CMP_AND_SWP;
+                send->s_wr.wr.atomic.compare_add = op->op_m_cswp.compare;
+                send->s_wr.wr.atomic.swap = op->op_m_cswp.swap;
+                send->s_wr.wr.atomic.compare_add_mask = op->op_m_cswp.compare_mask;
+                send->s_wr.wr.atomic.swap_mask = op->op_m_cswp.swap_mask;
+        } else { /* FADD */
+                send->s_wr.opcode = IB_WR_MASKED_ATOMIC_FETCH_AND_ADD;
+                send->s_wr.wr.atomic.compare_add = op->op_m_fadd.add;
+                send->s_wr.wr.atomic.swap = 0;
+                send->s_wr.wr.atomic.compare_add_mask = op->op_m_fadd.nocarry_mask;
+                send->s_wr.wr.atomic.swap_mask = 0;
+        }
+        nr_sig = rds_ib_set_wr_signal_state(ic, send, op->op_notify);
+        send->s_wr.num_sge = 1;
+        send->s_wr.next = NULL;
+        send->s_wr.wr.atomic.remote_addr = op->op_remote_addr;
+        send->s_wr.wr.atomic.rkey = op->op_rkey;
+        send->s_op = op;
+        rds_message_addref(container_of(send->s_op, struct rds_message, atomic));
+        /* map 8 byte retval buffer to the device */
+        ret = ib_dma_map_sg(ic->i_cm_id->device, op->op_sg, 1, DMA_FROM_DEVICE);
+        rdsdebug("ic %p mapping atomic op %p. mapped %d pg\n", ic, op, ret);
+        if (ret != 1) {
+                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
+                rds_ib_stats_inc(s_ib_tx_sg_mapping_failure);
+                ret = -ENOMEM; /* XXX ? */
+                goto out;
+        }
+        /* Convert our struct scatterlist to struct ib_sge */
+        send->s_sge[0].addr = ib_sg_dma_address(ic->i_cm_id->device, op->op_sg);
+        send->s_sge[0].length = ib_sg_dma_len(ic->i_cm_id->device, op->op_sg);
+        send->s_sge[0].lkey = ic->i_mr->lkey;
+        rdsdebug("rva %Lx rpa %Lx len %u\n", op->op_remote_addr,
+                 send->s_sge[0].addr, send->s_sge[0].length);
+        if (nr_sig)
+                atomic_add(nr_sig, &ic->i_signaled_sends);
+        failed_wr = &send->s_wr;
+        ret = ib_post_send(ic->i_cm_id->qp, &send->s_wr, &failed_wr);
+        rdsdebug("ic %p send %p (wr %p) ret %d wr %p\n", ic,
+                 send, &send->s_wr, ret, failed_wr);
+        BUG_ON(failed_wr != &send->s_wr);
+        if (ret) {
+                printk(KERN_WARNING "RDS/IB: atomic ib_post_send to %pI4 "
+                       "returned %d\n", &conn->c_faddr, ret);
+                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
+                rds_ib_sub_signaled(ic, nr_sig);
+                goto out;
+        }
+        if (unlikely(failed_wr != &send->s_wr)) {
+                printk(KERN_WARNING "RDS/IB: atomic ib_post_send() rc=%d, but failed_wqe updated!\n", ret);
+                BUG_ON(failed_wr != &send->s_wr);
+        }
+out:
+        return ret;
+}
+int rds_ib_xmit_rdma(struct rds_connection *conn, struct rm_rdma_op *op)
 {
        struct rds_ib_connection *ic = conn->c_transport_data;
        struct rds_ib_send_work *send = NULL;
        struct rds_ib_send_work *first;
        struct rds_ib_send_work *prev;
        struct ib_send_wr *failed_wr;
-        struct rds_ib_device *rds_ibdev;
        struct scatterlist *scat;
        unsigned long len;
-        u64 remote_addr = op->r_remote_addr;
+        u64 remote_addr = op->op_remote_addr;
+        u32 max_sge = ic->rds_ibdev->max_sge;
        u32 pos;
        u32 work_alloc;
        u32 i;
@@ -753,29 +888,28 @@ int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        int sent;
        int ret;
        int num_sge;
+        int nr_sig = 0;
-        rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
+        /* map the op the first time we see it */
-        /* map the message the first time we see it */
+        if (!op->op_mapped) {
-        if (!op->r_mapped) {
+                op->op_count = ib_dma_map_sg(ic->i_cm_id->device,
-                op->r_count = ib_dma_map_sg(ic->i_cm_id->device,
+                                             op->op_sg, op->op_nents, (op->op_write) ?
-                                        op->r_sg, op->r_nents, (op->r_write) ?
+                                             DMA_TO_DEVICE : DMA_FROM_DEVICE);
-                                        DMA_TO_DEVICE : DMA_FROM_DEVICE);
+                rdsdebug("ic %p mapping op %p: %d\n", ic, op, op->op_count);
-                rdsdebug("ic %p mapping op %p: %d\n", ic, op, op->r_count);
+                if (op->op_count == 0) {
-                if (op->r_count == 0) {
                        rds_ib_stats_inc(s_ib_tx_sg_mapping_failure);
                        ret = -ENOMEM; /* XXX ? */
                        goto out;
                }
-                op->r_mapped = 1;
+                op->op_mapped = 1;
        }
        /*
         * Instead of knowing how to return a partial rdma read/write we insist that there
         * be enough work requests to send the entire message.
         */
-        i = ceil(op->r_count, rds_ibdev->max_sge);
+        i = ceil(op->op_count, max_sge);
        work_alloc = rds_ib_ring_alloc(&ic->i_send_ring, i, &pos);
        if (work_alloc != i) {
@@ -788,30 +922,24 @@ int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        send = &ic->i_sends[pos];
        first = send;
        prev = NULL;
-        scat = &op->r_sg[0];
+        scat = &op->op_sg[0];
        sent = 0;
-        num_sge = op->r_count;
+        num_sge = op->op_count;
-        for (i = 0; i < work_alloc && scat != &op->r_sg[op->r_count]; i++) {
+        for (i = 0; i < work_alloc && scat != &op->op_sg[op->op_count]; i++) {
                send->s_wr.send_flags = 0;
                send->s_queued = jiffies;
-                /*
+                send->s_op = NULL;
-                 * We want to delay signaling completions just enough to get
-                 * the batching benefits but not so much that we create dead time on the wire.
+                nr_sig += rds_ib_set_wr_signal_state(ic, send, op->op_notify);
-                 */
-                if (ic->i_unsignaled_wrs-- == 0) {
-                        ic->i_unsignaled_wrs = rds_ib_sysctl_max_unsig_wrs;
-                        send->s_wr.send_flags = IB_SEND_SIGNALED;
-                }
-                send->s_wr.opcode = op->r_write ? IB_WR_RDMA_WRITE : IB_WR_RDMA_READ;
+                send->s_wr.opcode = op->op_write ? IB_WR_RDMA_WRITE : IB_WR_RDMA_READ;
                send->s_wr.wr.rdma.remote_addr = remote_addr;
-                send->s_wr.wr.rdma.rkey = op->r_key;
+                send->s_wr.wr.rdma.rkey = op->op_rkey;
-                send->s_op = op;
-                if (num_sge > rds_ibdev->max_sge) {
+                if (num_sge > max_sge) {
-                        send->s_wr.num_sge = rds_ibdev->max_sge;
+                        send->s_wr.num_sge = max_sge;
-                        num_sge -= rds_ibdev->max_sge;
+                        num_sge -= max_sge;
                } else {
                        send->s_wr.num_sge = num_sge;
                }
@@ -821,7 +949,7 @@ int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
                if (prev)
                        prev->s_wr.next = &send->s_wr;
-                for (j = 0; j < send->s_wr.num_sge && scat != &op->r_sg[op->r_count]; j++) {
+                for (j = 0; j < send->s_wr.num_sge && scat != &op->op_sg[op->op_count]; j++) {
                        len = ib_sg_dma_len(ic->i_cm_id->device, scat);
                        send->s_sge[j].addr =
                                 ib_sg_dma_address(ic->i_cm_id->device, scat);
@@ -843,15 +971,20 @@ int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
                        send = ic->i_sends;
        }
-        /* if we finished the message then send completion owns it */
+        /* give a reference to the last op */
-        if (scat == &op->r_sg[op->r_count])
+        if (scat == &op->op_sg[op->op_count]) {
-                prev->s_wr.send_flags = IB_SEND_SIGNALED;
+                prev->s_op = op;
+                rds_message_addref(container_of(op, struct rds_message, rdma));
+        }
        if (i < work_alloc) {
                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc - i);
                work_alloc = i;
        }
+        if (nr_sig)
+                atomic_add(nr_sig, &ic->i_signaled_sends);
        failed_wr = &first->s_wr;
        ret = ib_post_send(ic->i_cm_id->qp, &first->s_wr, &failed_wr);
        rdsdebug("ic %p first %p (wr %p) ret %d wr %p\n", ic,
@@ -861,6 +994,7 @@ int rds_ib_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
                printk(KERN_WARNING "RDS/IB: rdma ib_post_send to %pI4 "
                       "returned %d\n", &conn->c_faddr, ret);
                rds_ib_ring_unalloc(&ic->i_send_ring, work_alloc);
+                rds_ib_sub_signaled(ic, nr_sig);
                goto out;
        }
diff --git a/net/rds/ib_stats.c b/net/rds/ib_stats.c
index d2c904dd6fbc..2d5965d6e97c 100644
--- a/net/rds/ib_stats.c
+++ b/net/rds/ib_stats.c
@@ -67,6 +67,8 @@ static const char *const rds_ib_stat_names[] = {
        "ib_rdma_mr_pool_flush",
        "ib_rdma_mr_pool_wait",
        "ib_rdma_mr_pool_depleted",
+        "ib_atomic_cswp",
+        "ib_atomic_fadd",
 };
 unsigned int rds_ib_stats_info_copy(struct rds_info_iterator *iter,
diff --git a/net/rds/ib_sysctl.c b/net/rds/ib_sysctl.c
index 03f01cb4e0fe..fc3da37220fd 100644
--- a/net/rds/ib_sysctl.c
+++ b/net/rds/ib_sysctl.c
@@ -49,10 +49,6 @@ unsigned long rds_ib_sysctl_max_unsig_wrs = 16;
 static unsigned long rds_ib_sysctl_max_unsig_wr_min = 1;
 static unsigned long rds_ib_sysctl_max_unsig_wr_max = 64;
-unsigned long rds_ib_sysctl_max_unsig_bytes = (16 << 20);
-static unsigned long rds_ib_sysctl_max_unsig_bytes_min = 1;
-static unsigned long rds_ib_sysctl_max_unsig_bytes_max = ~0UL;
 /*
 * This sysctl does nothing.
 *
@@ -94,15 +90,6 @@ ctl_table rds_ib_sysctl_table[] = {
                .extra2         = &rds_ib_sysctl_max_unsig_wr_max,
        },
        {
-                .procname       = "max_unsignaled_bytes",
-                .data           = &rds_ib_sysctl_max_unsig_bytes,
-                .maxlen         = sizeof(unsigned long),
-                .mode           = 0644,
-                .proc_handler   = proc_doulongvec_minmax,
-                .extra1         = &rds_ib_sysctl_max_unsig_bytes_min,
-                .extra2         = &rds_ib_sysctl_max_unsig_bytes_max,
-        },
-        {
                .procname       = "max_recv_allocation",
                .data           = &rds_ib_sysctl_max_recv_allocation,
                .maxlen         = sizeof(unsigned long),
@@ -132,10 +119,10 @@ void rds_ib_sysctl_exit(void)
                unregister_sysctl_table(rds_ib_sysctl_hdr);
 }
-int __init rds_ib_sysctl_init(void)
+int rds_ib_sysctl_init(void)
 {
        rds_ib_sysctl_hdr = register_sysctl_paths(rds_ib_sysctl_path, rds_ib_sysctl_table);
-        if (rds_ib_sysctl_hdr == NULL)
+        if (!rds_ib_sysctl_hdr)
                return -ENOMEM;
        return 0;
 }
diff --git a/net/rds/info.c b/net/rds/info.c
index c45c4173a44d..4fdf1b6e84ff 100644
--- a/net/rds/info.c
+++ b/net/rds/info.c
@@ -76,7 +76,7 @@ void rds_info_register_func(int optname, rds_info_func func)
        BUG_ON(optname < RDS_INFO_FIRST || optname > RDS_INFO_LAST);
        spin_lock(&rds_info_lock);
-        BUG_ON(rds_info_funcs[offset] != NULL);
+        BUG_ON(rds_info_funcs[offset]);
        rds_info_funcs[offset] = func;
        spin_unlock(&rds_info_lock);
 }
@@ -102,7 +102,7 @@ EXPORT_SYMBOL_GPL(rds_info_deregister_func);
 */
 void rds_info_iter_unmap(struct rds_info_iterator *iter)
 {
-        if (iter->addr != NULL) {
+        if (iter->addr) {
                kunmap_atomic(iter->addr, KM_USER0);
                iter->addr = NULL;
        }
@@ -117,7 +117,7 @@ void rds_info_copy(struct rds_info_iterator *iter, void *data,
        unsigned long this;
        while (bytes) {
-                if (iter->addr == NULL)
+                if (!iter->addr)
                        iter->addr = kmap_atomic(*iter->pages, KM_USER0);
                this = min(bytes, PAGE_SIZE - iter->offset);
@@ -188,7 +188,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
                        >> PAGE_SHIFT;
        pages = kmalloc(nr_pages * sizeof(struct page *), GFP_KERNEL);
-        if (pages == NULL) {
+        if (!pages) {
                ret = -ENOMEM;
                goto out;
        }
@@ -206,7 +206,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 call_func:
        func = rds_info_funcs[optname - RDS_INFO_FIRST];
-        if (func == NULL) {
+        if (!func) {
                ret = -ENOPROTOOPT;
                goto out;
        }
@@ -234,7 +234,7 @@ call_func:
                ret = -EFAULT;
 out:
-        for (i = 0; pages != NULL && i < nr_pages; i++)
+        for (i = 0; pages && i < nr_pages; i++)
                put_page(pages[i]);
        kfree(pages);
diff --git a/net/rds/iw.c b/net/rds/iw.c
index c8f3d3525cb9..56808cac0fc7 100644
--- a/net/rds/iw.c
+++ b/net/rds/iw.c
@@ -264,7 +264,6 @@ struct rds_transport rds_iw_transport = {
        .laddr_check            = rds_iw_laddr_check,
        .xmit_complete          = rds_iw_xmit_complete,
        .xmit                   = rds_iw_xmit,
-        .xmit_cong_map          = NULL,
        .xmit_rdma              = rds_iw_xmit_rdma,
        .recv                   = rds_iw_recv,
        .conn_alloc             = rds_iw_conn_alloc,
@@ -272,7 +271,6 @@ struct rds_transport rds_iw_transport = {
        .conn_connect           = rds_iw_conn_connect,
        .conn_shutdown          = rds_iw_conn_shutdown,
        .inc_copy_to_user       = rds_iw_inc_copy_to_user,
-        .inc_purge              = rds_iw_inc_purge,
        .inc_free               = rds_iw_inc_free,
        .cm_initiate_connect    = rds_iw_cm_initiate_connect,
        .cm_handle_connect      = rds_iw_cm_handle_connect,
@@ -289,7 +287,7 @@ struct rds_transport rds_iw_transport = {
        .t_prefer_loopback      = 1,
 };
-int __init rds_iw_init(void)
+int rds_iw_init(void)
 {
        int ret;
diff --git a/net/rds/iw.h b/net/rds/iw.h
index eef2f0c28476..543e665fafe3 100644
--- a/net/rds/iw.h
+++ b/net/rds/iw.h
@@ -70,7 +70,7 @@ struct rds_iw_send_work {
        struct rds_message      *s_rm;
        /* We should really put these into a union: */
-        struct rds_rdma_op      *s_op;
+        struct rm_rdma_op       *s_op;
        struct rds_iw_mapping   *s_mapping;
        struct ib_mr            *s_mr;
        struct ib_fast_reg_page_list *s_page_list;
@@ -284,7 +284,7 @@ void rds_iw_conn_free(void *arg);
 int rds_iw_conn_connect(struct rds_connection *conn);
 void rds_iw_conn_shutdown(struct rds_connection *conn);
 void rds_iw_state_change(struct sock *sk);
-int __init rds_iw_listen_init(void);
+int rds_iw_listen_init(void);
 void rds_iw_listen_stop(void);
 void __rds_iw_conn_error(struct rds_connection *conn, const char *, ...);
 int rds_iw_cm_handle_connect(struct rdma_cm_id *cm_id,
@@ -321,12 +321,11 @@ void rds_iw_flush_mrs(void);
 void rds_iw_remove_cm_id(struct rds_iw_device *rds_iwdev, struct rdma_cm_id *cm_id);
 /* ib_recv.c */
-int __init rds_iw_recv_init(void);
+int rds_iw_recv_init(void);
 void rds_iw_recv_exit(void);
 int rds_iw_recv(struct rds_connection *conn);
 int rds_iw_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
                       gfp_t page_gfp, int prefill);
-void rds_iw_inc_purge(struct rds_incoming *inc);
 void rds_iw_inc_free(struct rds_incoming *inc);
 int rds_iw_inc_copy_to_user(struct rds_incoming *inc, struct iovec *iov,
                             size_t size);
@@ -358,7 +357,7 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
 void rds_iw_send_cq_comp_handler(struct ib_cq *cq, void *context);
 void rds_iw_send_init_ring(struct rds_iw_connection *ic);
 void rds_iw_send_clear_ring(struct rds_iw_connection *ic);
-int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op);
+int rds_iw_xmit_rdma(struct rds_connection *conn, struct rm_rdma_op *op);
 void rds_iw_send_add_credits(struct rds_connection *conn, unsigned int credits);
 void rds_iw_advertise_credits(struct rds_connection *conn, unsigned int posted);
 int rds_iw_send_grab_credits(struct rds_iw_connection *ic, u32 wanted,
@@ -371,7 +370,7 @@ unsigned int rds_iw_stats_info_copy(struct rds_info_iterator *iter,
                                    unsigned int avail);
 /* ib_sysctl.c */
-int __init rds_iw_sysctl_init(void);
+int rds_iw_sysctl_init(void);
 void rds_iw_sysctl_exit(void);
 extern unsigned long rds_iw_sysctl_max_send_wr;
 extern unsigned long rds_iw_sysctl_max_recv_wr;
diff --git a/net/rds/iw_cm.c b/net/rds/iw_cm.c
index b5dd6ac39be8..712cf2d1f28e 100644
--- a/net/rds/iw_cm.c
+++ b/net/rds/iw_cm.c
@@ -257,7 +257,7 @@ static int rds_iw_setup_qp(struct rds_connection *conn)
         * the rds_iwdev at all.
         */
        rds_iwdev = ib_get_client_data(dev, &rds_iw_client);
-        if (rds_iwdev == NULL) {
+        if (!rds_iwdev) {
                if (printk_ratelimit())
                        printk(KERN_NOTICE "RDS/IW: No client_data for device %s\n",
                                        dev->name);
@@ -292,7 +292,7 @@ static int rds_iw_setup_qp(struct rds_connection *conn)
                                           ic->i_send_ring.w_nr *
                                                sizeof(struct rds_header),
                                           &ic->i_send_hdrs_dma, GFP_KERNEL);
-        if (ic->i_send_hdrs == NULL) {
+        if (!ic->i_send_hdrs) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent send failed\n");
                goto out;
@@ -302,7 +302,7 @@ static int rds_iw_setup_qp(struct rds_connection *conn)
                                           ic->i_recv_ring.w_nr *
                                                sizeof(struct rds_header),
                                           &ic->i_recv_hdrs_dma, GFP_KERNEL);
-        if (ic->i_recv_hdrs == NULL) {
+        if (!ic->i_recv_hdrs) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent recv failed\n");
                goto out;
@@ -310,14 +310,14 @@ static int rds_iw_setup_qp(struct rds_connection *conn)
        ic->i_ack = ib_dma_alloc_coherent(dev, sizeof(struct rds_header),
                                       &ic->i_ack_dma, GFP_KERNEL);
-        if (ic->i_ack == NULL) {
+        if (!ic->i_ack) {
                ret = -ENOMEM;
                rdsdebug("ib_dma_alloc_coherent ack failed\n");
                goto out;
        }
        ic->i_sends = vmalloc(ic->i_send_ring.w_nr * sizeof(struct rds_iw_send_work));
-        if (ic->i_sends == NULL) {
+        if (!ic->i_sends) {
                ret = -ENOMEM;
                rdsdebug("send allocation failed\n");
                goto out;
@@ -325,7 +325,7 @@ static int rds_iw_setup_qp(struct rds_connection *conn)
        rds_iw_send_init_ring(ic);
        ic->i_recvs = vmalloc(ic->i_recv_ring.w_nr * sizeof(struct rds_iw_recv_work));
-        if (ic->i_recvs == NULL) {
+        if (!ic->i_recvs) {
                ret = -ENOMEM;
                rdsdebug("recv allocation failed\n");
                goto out;
@@ -696,7 +696,7 @@ int rds_iw_conn_alloc(struct rds_connection *conn, gfp_t gfp)
        /* XXX too lazy? */
        ic = kzalloc(sizeof(struct rds_iw_connection), GFP_KERNEL);
-        if (ic == NULL)
+        if (!ic)
                return -ENOMEM;
        INIT_LIST_HEAD(&ic->iw_node);
diff --git a/net/rds/iw_rdma.c b/net/rds/iw_rdma.c
index 13dc1862d862..0e7accc23ee2 100644
--- a/net/rds/iw_rdma.c
+++ b/net/rds/iw_rdma.c
@@ -34,7 +34,6 @@
 #include <linux/slab.h>
 #include "rds.h"
-#include "rdma.h"
 #include "iw.h"
@@ -207,9 +206,9 @@ void rds_iw_add_conn(struct rds_iw_device *rds_iwdev, struct rds_connection *con
        BUG_ON(list_empty(&ic->iw_node));
        list_del(&ic->iw_node);
-        spin_lock_irq(&rds_iwdev->spinlock);
+        spin_lock(&rds_iwdev->spinlock);
        list_add_tail(&ic->iw_node, &rds_iwdev->conn_list);
-        spin_unlock_irq(&rds_iwdev->spinlock);
+        spin_unlock(&rds_iwdev->spinlock);
        spin_unlock_irq(&iw_nodev_conns_lock);
        ic->rds_iwdev = rds_iwdev;
diff --git a/net/rds/iw_recv.c b/net/rds/iw_recv.c
index 3d479067d54d..5e57347f49ff 100644
--- a/net/rds/iw_recv.c
+++ b/net/rds/iw_recv.c
@@ -53,7 +53,7 @@ static void rds_iw_frag_drop_page(struct rds_page_frag *frag)
 static void rds_iw_frag_free(struct rds_page_frag *frag)
 {
        rdsdebug("frag %p page %p\n", frag, frag->f_page);
-        BUG_ON(frag->f_page != NULL);
+        BUG_ON(frag->f_page);
        kmem_cache_free(rds_iw_frag_slab, frag);
 }
@@ -143,14 +143,14 @@ static int rds_iw_recv_refill_one(struct rds_connection *conn,
        struct ib_sge *sge;
        int ret = -ENOMEM;
-        if (recv->r_iwinc == NULL) {
+        if (!recv->r_iwinc) {
                if (!atomic_add_unless(&rds_iw_allocation, 1, rds_iw_sysctl_max_recv_allocation)) {
                        rds_iw_stats_inc(s_iw_rx_alloc_limit);
                        goto out;
                }
                recv->r_iwinc = kmem_cache_alloc(rds_iw_incoming_slab,
                                                 kptr_gfp);
-                if (recv->r_iwinc == NULL) {
+                if (!recv->r_iwinc) {
                        atomic_dec(&rds_iw_allocation);
                        goto out;
                }
@@ -158,17 +158,17 @@ static int rds_iw_recv_refill_one(struct rds_connection *conn,
                rds_inc_init(&recv->r_iwinc->ii_inc, conn, conn->c_faddr);
        }
-        if (recv->r_frag == NULL) {
+        if (!recv->r_frag) {
                recv->r_frag = kmem_cache_alloc(rds_iw_frag_slab, kptr_gfp);
-                if (recv->r_frag == NULL)
+                if (!recv->r_frag)
                        goto out;
                INIT_LIST_HEAD(&recv->r_frag->f_item);
                recv->r_frag->f_page = NULL;
        }
-        if (ic->i_frag.f_page == NULL) {
+        if (!ic->i_frag.f_page) {
                ic->i_frag.f_page = alloc_page(page_gfp);
-                if (ic->i_frag.f_page == NULL)
+                if (!ic->i_frag.f_page)
                        goto out;
                ic->i_frag.f_offset = 0;
        }
@@ -273,7 +273,7 @@ int rds_iw_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
        return ret;
 }
-void rds_iw_inc_purge(struct rds_incoming *inc)
+static void rds_iw_inc_purge(struct rds_incoming *inc)
 {
        struct rds_iw_incoming *iwinc;
        struct rds_page_frag *frag;
@@ -716,7 +716,7 @@ static void rds_iw_process_recv(struct rds_connection *conn,
         * into the inc and save the inc so we can hang upcoming fragments
         * off its list.
         */
-        if (iwinc == NULL) {
+        if (!iwinc) {
                iwinc = recv->r_iwinc;
                recv->r_iwinc = NULL;
                ic->i_iwinc = iwinc;
@@ -887,7 +887,7 @@ int rds_iw_recv(struct rds_connection *conn)
        return ret;
 }
-int __init rds_iw_recv_init(void)
+int rds_iw_recv_init(void)
 {
        struct sysinfo si;
        int ret = -ENOMEM;
@@ -899,13 +899,13 @@ int __init rds_iw_recv_init(void)
        rds_iw_incoming_slab = kmem_cache_create("rds_iw_incoming",
                                        sizeof(struct rds_iw_incoming),
                                        0, 0, NULL);
-        if (rds_iw_incoming_slab == NULL)
+        if (!rds_iw_incoming_slab)
                goto out;
        rds_iw_frag_slab = kmem_cache_create("rds_iw_frag",
                                        sizeof(struct rds_page_frag),
                                        0, 0, NULL);
-        if (rds_iw_frag_slab == NULL)
+        if (!rds_iw_frag_slab)
                kmem_cache_destroy(rds_iw_incoming_slab);
        else
                ret = 0;
diff --git a/net/rds/iw_send.c b/net/rds/iw_send.c
index 52182ff7519e..6280ea020d4e 100644
--- a/net/rds/iw_send.c
+++ b/net/rds/iw_send.c
@@ -36,7 +36,6 @@
 #include <linux/dmapool.h>
 #include "rds.h"
-#include "rdma.h"
 #include "iw.h"
 static void rds_iw_send_rdma_complete(struct rds_message *rm,
@@ -64,13 +63,13 @@ static void rds_iw_send_rdma_complete(struct rds_message *rm,
 }
 static void rds_iw_send_unmap_rdma(struct rds_iw_connection *ic,
-                                   struct rds_rdma_op *op)
+                                   struct rm_rdma_op *op)
 {
-        if (op->r_mapped) {
+        if (op->op_mapped) {
                ib_dma_unmap_sg(ic->i_cm_id->device,
-                        op->r_sg, op->r_nents,
+                        op->op_sg, op->op_nents,
-                        op->r_write ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
+                        op->op_write ? DMA_TO_DEVICE : DMA_FROM_DEVICE);
-                op->r_mapped = 0;
+                op->op_mapped = 0;
        }
 }
@@ -83,11 +82,11 @@ static void rds_iw_send_unmap_rm(struct rds_iw_connection *ic,
        rdsdebug("ic %p send %p rm %p\n", ic, send, rm);
        ib_dma_unmap_sg(ic->i_cm_id->device,
-                     rm->m_sg, rm->m_nents,
+                     rm->data.op_sg, rm->data.op_nents,
                     DMA_TO_DEVICE);
-        if (rm->m_rdma_op != NULL) {
+        if (rm->rdma.op_active) {
-                rds_iw_send_unmap_rdma(ic, rm->m_rdma_op);
+                rds_iw_send_unmap_rdma(ic, &rm->rdma);
                /* If the user asked for a completion notification on this
                 * message, we can implement three different semantics:
@@ -111,10 +110,10 @@ static void rds_iw_send_unmap_rm(struct rds_iw_connection *ic,
                 */
                rds_iw_send_rdma_complete(rm, wc_status);
-                if (rm->m_rdma_op->r_write)
+                if (rm->rdma.op_write)
-                        rds_stats_add(s_send_rdma_bytes, rm->m_rdma_op->r_bytes);
+                        rds_stats_add(s_send_rdma_bytes, rm->rdma.op_bytes);
                else
-                        rds_stats_add(s_recv_rdma_bytes, rm->m_rdma_op->r_bytes);
+                        rds_stats_add(s_recv_rdma_bytes, rm->rdma.op_bytes);
        }
        /* If anyone waited for this message to get flushed out, wake
@@ -556,25 +555,27 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
        }
        /* map the message the first time we see it */
-        if (ic->i_rm == NULL) {
+        if (!ic->i_rm) {
                /*
                printk(KERN_NOTICE "rds_iw_xmit prep msg dport=%u flags=0x%x len=%d\n",
                                be16_to_cpu(rm->m_inc.i_hdr.h_dport),
                                rm->m_inc.i_hdr.h_flags,
                                be32_to_cpu(rm->m_inc.i_hdr.h_len));
                   */
-                if (rm->m_nents) {
+                if (rm->data.op_nents) {
-                        rm->m_count = ib_dma_map_sg(dev,
+                        rm->data.op_count = ib_dma_map_sg(dev,
-                                         rm->m_sg, rm->m_nents, DMA_TO_DEVICE);
+                                                          rm->data.op_sg,
-                        rdsdebug("ic %p mapping rm %p: %d\n", ic, rm, rm->m_count);
+                                                          rm->data.op_nents,
-                        if (rm->m_count == 0) {
+                                                          DMA_TO_DEVICE);
+                        rdsdebug("ic %p mapping rm %p: %d\n", ic, rm, rm->data.op_count);
+                        if (rm->data.op_count == 0) {
                                rds_iw_stats_inc(s_iw_tx_sg_mapping_failure);
                                rds_iw_ring_unalloc(&ic->i_send_ring, work_alloc);
                                ret = -ENOMEM; /* XXX ? */
                                goto out;
                        }
                } else {
-                        rm->m_count = 0;
+                        rm->data.op_count = 0;
                }
                ic->i_unsignaled_wrs = rds_iw_sysctl_max_unsig_wrs;
@@ -590,10 +591,10 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
                /* If it has a RDMA op, tell the peer we did it. This is
                 * used by the peer to release use-once RDMA MRs. */
-                if (rm->m_rdma_op) {
+                if (rm->rdma.op_active) {
                        struct rds_ext_header_rdma ext_hdr;
-                        ext_hdr.h_rdma_rkey = cpu_to_be32(rm->m_rdma_op->r_key);
+                        ext_hdr.h_rdma_rkey = cpu_to_be32(rm->rdma.op_rkey);
                        rds_message_add_extension(&rm->m_inc.i_hdr,
                                        RDS_EXTHDR_RDMA, &ext_hdr, sizeof(ext_hdr));
                }
@@ -621,7 +622,7 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
        send = &ic->i_sends[pos];
        first = send;
        prev = NULL;
-        scat = &rm->m_sg[sg];
+        scat = &rm->data.op_sg[sg];
        sent = 0;
        i = 0;
@@ -631,7 +632,7 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
         * or when requested by the user. Right now, we let
         * the application choose.
         */
-        if (rm->m_rdma_op && rm->m_rdma_op->r_fence)
+        if (rm->rdma.op_active && rm->rdma.op_fence)
                send_flags = IB_SEND_FENCE;
        /*
@@ -650,7 +651,7 @@ int rds_iw_xmit(struct rds_connection *conn, struct rds_message *rm,
        }
        /* if there's data reference it with a chain of work reqs */
-        for (; i < work_alloc && scat != &rm->m_sg[rm->m_count]; i++) {
+        for (; i < work_alloc && scat != &rm->data.op_sg[rm->data.op_count]; i++) {
                unsigned int len;
                send = &ic->i_sends[pos];
@@ -728,7 +729,7 @@ add_header:
                sent += sizeof(struct rds_header);
        /* if we finished the message then send completion owns it */
-        if (scat == &rm->m_sg[rm->m_count]) {
+        if (scat == &rm->data.op_sg[rm->data.op_count]) {
                prev->s_rm = ic->i_rm;
                prev->s_wr.send_flags |= IB_SEND_SIGNALED | IB_SEND_SOLICITED;
                ic->i_rm = NULL;
@@ -784,7 +785,7 @@ static void rds_iw_build_send_fastreg(struct rds_iw_device *rds_iwdev, struct rd
        ib_update_fast_reg_key(send->s_mr, send->s_remap_count++);
 }
-int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
+int rds_iw_xmit_rdma(struct rds_connection *conn, struct rm_rdma_op *op)
 {
        struct rds_iw_connection *ic = conn->c_transport_data;
        struct rds_iw_send_work *send = NULL;
@@ -794,7 +795,7 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        struct rds_iw_device *rds_iwdev;
        struct scatterlist *scat;
        unsigned long len;
-        u64 remote_addr = op->r_remote_addr;
+        u64 remote_addr = op->op_remote_addr;
        u32 pos, fr_pos;
        u32 work_alloc;
        u32 i;
@@ -806,21 +807,21 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        rds_iwdev = ib_get_client_data(ic->i_cm_id->device, &rds_iw_client);
        /* map the message the first time we see it */
-        if (!op->r_mapped) {
+        if (!op->op_mapped) {
-                op->r_count = ib_dma_map_sg(ic->i_cm_id->device,
+                op->op_count = ib_dma_map_sg(ic->i_cm_id->device,
-                                        op->r_sg, op->r_nents, (op->r_write) ?
+                                             op->op_sg, op->op_nents, (op->op_write) ?
-                                        DMA_TO_DEVICE : DMA_FROM_DEVICE);
+                                             DMA_TO_DEVICE : DMA_FROM_DEVICE);
-                rdsdebug("ic %p mapping op %p: %d\n", ic, op, op->r_count);
+                rdsdebug("ic %p mapping op %p: %d\n", ic, op, op->op_count);
-                if (op->r_count == 0) {
+                if (op->op_count == 0) {
                        rds_iw_stats_inc(s_iw_tx_sg_mapping_failure);
                        ret = -ENOMEM; /* XXX ? */
                        goto out;
                }
-                op->r_mapped = 1;
+                op->op_mapped = 1;
        }
-        if (!op->r_write) {
+        if (!op->op_write) {
                /* Alloc space on the send queue for the fastreg */
                work_alloc = rds_iw_ring_alloc(&ic->i_send_ring, 1, &fr_pos);
                if (work_alloc != 1) {
@@ -835,7 +836,7 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
         * Instead of knowing how to return a partial rdma read/write we insist that there
         * be enough work requests to send the entire message.
         */
-        i = ceil(op->r_count, rds_iwdev->max_sge);
+        i = ceil(op->op_count, rds_iwdev->max_sge);
        work_alloc = rds_iw_ring_alloc(&ic->i_send_ring, i, &pos);
        if (work_alloc != i) {
@@ -846,17 +847,17 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        }
        send = &ic->i_sends[pos];
-        if (!op->r_write) {
+        if (!op->op_write) {
                first = prev = &ic->i_sends[fr_pos];
        } else {
                first = send;
                prev = NULL;
        }
-        scat = &op->r_sg[0];
+        scat = &op->op_sg[0];
        sent = 0;
-        num_sge = op->r_count;
+        num_sge = op->op_count;
-        for (i = 0; i < work_alloc && scat != &op->r_sg[op->r_count]; i++) {
+        for (i = 0; i < work_alloc && scat != &op->op_sg[op->op_count]; i++) {
                send->s_wr.send_flags = 0;
                send->s_queued = jiffies;
@@ -873,13 +874,13 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
                 * for local access after RDS is finished with it, using
                 * IB_WR_RDMA_READ_WITH_INV will invalidate it after the read has completed.
                 */
-                if (op->r_write)
+                if (op->op_write)
                        send->s_wr.opcode = IB_WR_RDMA_WRITE;
                else
                        send->s_wr.opcode = IB_WR_RDMA_READ_WITH_INV;
                send->s_wr.wr.rdma.remote_addr = remote_addr;
-                send->s_wr.wr.rdma.rkey = op->r_key;
+                send->s_wr.wr.rdma.rkey = op->op_rkey;
                send->s_op = op;
                if (num_sge > rds_iwdev->max_sge) {
@@ -893,7 +894,7 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
                if (prev)
                        prev->s_wr.next = &send->s_wr;
-                for (j = 0; j < send->s_wr.num_sge && scat != &op->r_sg[op->r_count]; j++) {
+                for (j = 0; j < send->s_wr.num_sge && scat != &op->op_sg[op->op_count]; j++) {
                        len = ib_sg_dma_len(ic->i_cm_id->device, scat);
                        if (send->s_wr.opcode == IB_WR_RDMA_READ_WITH_INV)
@@ -927,7 +928,7 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
        }
        /* if we finished the message then send completion owns it */
-        if (scat == &op->r_sg[op->r_count])
+        if (scat == &op->op_sg[op->op_count])
                first->s_wr.send_flags = IB_SEND_SIGNALED;
        if (i < work_alloc) {
@@ -941,9 +942,9 @@ int rds_iw_xmit_rdma(struct rds_connection *conn, struct rds_rdma_op *op)
         * adapters do not allow using the lkey for this at all.  To bypass this use a
         * fastreg_mr (or possibly a dma_mr)
         */
-        if (!op->r_write) {
+        if (!op->op_write) {
                rds_iw_build_send_fastreg(rds_iwdev, ic, &ic->i_sends[fr_pos],
-                        op->r_count, sent, conn->c_xmit_rm->m_rs->rs_user_addr);
+                        op->op_count, sent, conn->c_xmit_rm->m_rs->rs_user_addr);
                work_alloc++;
        }
diff --git a/net/rds/iw_sysctl.c b/net/rds/iw_sysctl.c
index 1c4428a61a02..23e3a9a26aaf 100644
--- a/net/rds/iw_sysctl.c
+++ b/net/rds/iw_sysctl.c
@@ -122,10 +122,10 @@ void rds_iw_sysctl_exit(void)
                unregister_sysctl_table(rds_iw_sysctl_hdr);
 }
-int __init rds_iw_sysctl_init(void)
+int rds_iw_sysctl_init(void)
 {
        rds_iw_sysctl_hdr = register_sysctl_paths(rds_iw_sysctl_path, rds_iw_sysctl_table);
-        if (rds_iw_sysctl_hdr == NULL)
+        if (!rds_iw_sysctl_hdr)
                return -ENOMEM;
        return 0;
 }
diff --git a/net/rds/loop.c b/net/rds/loop.c
index dd9879379457..c390156b426f 100644
--- a/net/rds/loop.c
+++ b/net/rds/loop.c
@@ -61,10 +61,17 @@ static int rds_loop_xmit(struct rds_connection *conn, struct rds_message *rm,
                         unsigned int hdr_off, unsigned int sg,
                         unsigned int off)
 {
+        /* Do not send cong updates to loopback */
+        if (rm->m_inc.i_hdr.h_flags & RDS_FLAG_CONG_BITMAP) {
+                rds_cong_map_updated(conn->c_fcong, ~(u64) 0);
+                return sizeof(struct rds_header) + RDS_CONG_MAP_BYTES;
+        }
        BUG_ON(hdr_off || sg || off);
        rds_inc_init(&rm->m_inc, conn, conn->c_laddr);
-        rds_message_addref(rm); /* for the inc */
+        /* For the embedded inc. Matching put is in loop_inc_free() */
+        rds_message_addref(rm);
        rds_recv_incoming(conn, conn->c_laddr, conn->c_faddr, &rm->m_inc,
                          GFP_KERNEL, KM_USER0);
@@ -77,16 +84,14 @@ static int rds_loop_xmit(struct rds_connection *conn, struct rds_message *rm,
        return sizeof(struct rds_header) + be32_to_cpu(rm->m_inc.i_hdr.h_len);
 }
-static int rds_loop_xmit_cong_map(struct rds_connection *conn,
+/*
-                                  struct rds_cong_map *map,
+ * See rds_loop_xmit(). Since our inc is embedded in the rm, we
-                                  unsigned long offset)
+ * make sure the rm lives at least until the inc is done.
+ */
+static void rds_loop_inc_free(struct rds_incoming *inc)
 {
-        BUG_ON(offset);
+        struct rds_message *rm = container_of(inc, struct rds_message, m_inc);
-        BUG_ON(map != conn->c_lcong);
+        rds_message_put(rm);
-        rds_cong_map_updated(conn->c_fcong, ~(u64) 0);
-        return sizeof(struct rds_header) + RDS_CONG_MAP_BYTES;
 }
 /* we need to at least give the thread something to succeed */
@@ -112,7 +117,7 @@ static int rds_loop_conn_alloc(struct rds_connection *conn, gfp_t gfp)
        unsigned long flags;
        lc = kzalloc(sizeof(struct rds_loop_connection), GFP_KERNEL);
-        if (lc == NULL)
+        if (!lc)
                return -ENOMEM;
        INIT_LIST_HEAD(&lc->loop_node);
@@ -169,14 +174,12 @@ void rds_loop_exit(void)
 */
 struct rds_transport rds_loop_transport = {
        .xmit                   = rds_loop_xmit,
-        .xmit_cong_map          = rds_loop_xmit_cong_map,
        .recv                   = rds_loop_recv,
        .conn_alloc             = rds_loop_conn_alloc,
        .conn_free              = rds_loop_conn_free,
        .conn_connect           = rds_loop_conn_connect,
        .conn_shutdown          = rds_loop_conn_shutdown,
        .inc_copy_to_user       = rds_message_inc_copy_to_user,
-        .inc_purge              = rds_message_inc_purge,
+        .inc_free               = rds_loop_inc_free,
-        .inc_free               = rds_message_inc_free,
        .t_name                 = "loopback",
 };
diff --git a/net/rds/message.c b/net/rds/message.c
index 9a1d67e001ba..84f937f11d47 100644
--- a/net/rds/message.c
+++ b/net/rds/message.c
@@ -34,9 +34,6 @@
 #include <linux/slab.h>
 #include "rds.h"
-#include "rdma.h"
-static DECLARE_WAIT_QUEUE_HEAD(rds_message_flush_waitq);
 static unsigned int     rds_exthdr_size[__RDS_EXTHDR_MAX] = {
 [RDS_EXTHDR_NONE]       = 0,
@@ -63,29 +60,31 @@ static void rds_message_purge(struct rds_message *rm)
        if (unlikely(test_bit(RDS_MSG_PAGEVEC, &rm->m_flags)))
                return;
-        for (i = 0; i < rm->m_nents; i++) {
+        for (i = 0; i < rm->data.op_nents; i++) {
-                rdsdebug("putting data page %p\n", (void *)sg_page(&rm->m_sg[i]));
+                rdsdebug("putting data page %p\n", (void *)sg_page(&rm->data.op_sg[i]));
                /* XXX will have to put_page for page refs */
-                __free_page(sg_page(&rm->m_sg[i]));
+                __free_page(sg_page(&rm->data.op_sg[i]));
        }
-        rm->m_nents = 0;
+        rm->data.op_nents = 0;
-        if (rm->m_rdma_op)
+        if (rm->rdma.op_active)
-                rds_rdma_free_op(rm->m_rdma_op);
+                rds_rdma_free_op(&rm->rdma);
-        if (rm->m_rdma_mr)
+        if (rm->rdma.op_rdma_mr)
-                rds_mr_put(rm->m_rdma_mr);
+                rds_mr_put(rm->rdma.op_rdma_mr);
-}
-void rds_message_inc_purge(struct rds_incoming *inc)
+        if (rm->atomic.op_active)
-{
+                rds_atomic_free_op(&rm->atomic);
-        struct rds_message *rm = container_of(inc, struct rds_message, m_inc);
+        if (rm->atomic.op_rdma_mr)
-        rds_message_purge(rm);
+                rds_mr_put(rm->atomic.op_rdma_mr);
 }
 void rds_message_put(struct rds_message *rm)
 {
        rdsdebug("put rm %p ref %d\n", rm, atomic_read(&rm->m_refcount));
+        if (atomic_read(&rm->m_refcount) == 0) {
+printk(KERN_CRIT "danger refcount zero on %p\n", rm);
+WARN_ON(1);
+        }
        if (atomic_dec_and_test(&rm->m_refcount)) {
                BUG_ON(!list_empty(&rm->m_sock_item));
                BUG_ON(!list_empty(&rm->m_conn_item));
@@ -96,12 +95,6 @@ void rds_message_put(struct rds_message *rm)
 }
 EXPORT_SYMBOL_GPL(rds_message_put);
-void rds_message_inc_free(struct rds_incoming *inc)
-{
-        struct rds_message *rm = container_of(inc, struct rds_message, m_inc);
-        rds_message_put(rm);
-}
 void rds_message_populate_header(struct rds_header *hdr, __be16 sport,
                                 __be16 dport, u64 seq)
 {
@@ -214,41 +207,68 @@ int rds_message_add_rdma_dest_extension(struct rds_header *hdr, u32 r_key, u32 o
 }
 EXPORT_SYMBOL_GPL(rds_message_add_rdma_dest_extension);
-struct rds_message *rds_message_alloc(unsigned int nents, gfp_t gfp)
+/*
+ * Each rds_message is allocated with extra space for the scatterlist entries
+ * rds ops will need. This is to minimize memory allocation count. Then, each rds op
+ * can grab SGs when initializing its part of the rds_message.
+ */
+struct rds_message *rds_message_alloc(unsigned int extra_len, gfp_t gfp)
 {
        struct rds_message *rm;
-        rm = kzalloc(sizeof(struct rds_message) +
+        rm = kzalloc(sizeof(struct rds_message) + extra_len, gfp);
-                     (nents * sizeof(struct scatterlist)), gfp);
        if (!rm)
                goto out;
-        if (nents)
+        rm->m_used_sgs = 0;
-                sg_init_table(rm->m_sg, nents);
+        rm->m_total_sgs = extra_len / sizeof(struct scatterlist);
        atomic_set(&rm->m_refcount, 1);
        INIT_LIST_HEAD(&rm->m_sock_item);
        INIT_LIST_HEAD(&rm->m_conn_item);
        spin_lock_init(&rm->m_rs_lock);
+        init_waitqueue_head(&rm->m_flush_wait);
 out:
        return rm;
 }
+/*
+ * RDS ops use this to grab SG entries from the rm's sg pool.
+ */
+struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents)
+{
+        struct scatterlist *sg_first = (struct scatterlist *) &rm[1];
+        struct scatterlist *sg_ret;
+        WARN_ON(rm->m_used_sgs + nents > rm->m_total_sgs);
+        WARN_ON(!nents);
+        sg_ret = &sg_first[rm->m_used_sgs];
+        sg_init_table(sg_ret, nents);
+        rm->m_used_sgs += nents;
+        return sg_ret;
+}
 struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned int total_len)
 {
        struct rds_message *rm;
        unsigned int i;
+        int num_sgs = ceil(total_len, PAGE_SIZE);
+        int extra_bytes = num_sgs * sizeof(struct scatterlist);
-        rm = rds_message_alloc(ceil(total_len, PAGE_SIZE), GFP_KERNEL);
+        rm = rds_message_alloc(extra_bytes, GFP_NOWAIT);
-        if (rm == NULL)
+        if (!rm)
                return ERR_PTR(-ENOMEM);
        set_bit(RDS_MSG_PAGEVEC, &rm->m_flags);
        rm->m_inc.i_hdr.h_len = cpu_to_be32(total_len);
-        rm->m_nents = ceil(total_len, PAGE_SIZE);
+        rm->data.op_nents = ceil(total_len, PAGE_SIZE);
+        rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs);
-        for (i = 0; i < rm->m_nents; ++i) {
+        for (i = 0; i < rm->data.op_nents; ++i) {
-                sg_set_page(&rm->m_sg[i],
+                sg_set_page(&rm->data.op_sg[i],
                                virt_to_page(page_addrs[i]),
                                PAGE_SIZE, 0);
        }
@@ -256,40 +276,33 @@ struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned in
        return rm;
 }
-struct rds_message *rds_message_copy_from_user(struct iovec *first_iov,
+int rds_message_copy_from_user(struct rds_message *rm, struct iovec *first_iov,
                                               size_t total_len)
 {
        unsigned long to_copy;
        unsigned long iov_off;
        unsigned long sg_off;
-        struct rds_message *rm;
        struct iovec *iov;
        struct scatterlist *sg;
-        int ret;
+        int ret = 0;
-        rm = rds_message_alloc(ceil(total_len, PAGE_SIZE), GFP_KERNEL);
-        if (rm == NULL) {
-                ret = -ENOMEM;
-                goto out;
-        }
        rm->m_inc.i_hdr.h_len = cpu_to_be32(total_len);
        /*
         * now allocate and copy in the data payload.
         */
-        sg = rm->m_sg;
+        sg = rm->data.op_sg;
        iov = first_iov;
        iov_off = 0;
        sg_off = 0; /* Dear gcc, sg->page will be null from kzalloc. */
        while (total_len) {
-                if (sg_page(sg) == NULL) {
+                if (!sg_page(sg)) {
                        ret = rds_page_remainder_alloc(sg, total_len,
                                                       GFP_HIGHUSER);
                        if (ret)
                                goto out;
-                        rm->m_nents++;
+                        rm->data.op_nents++;
                        sg_off = 0;
                }
@@ -320,14 +333,8 @@ struct rds_message *rds_message_copy_from_user(struct iovec *first_iov,
                        sg++;
        }
-        ret = 0;
 out:
-        if (ret) {
+        return ret;
-                if (rm)
-                        rds_message_put(rm);
-                rm = ERR_PTR(ret);
-        }
-        return rm;
 }
 int rds_message_inc_copy_to_user(struct rds_incoming *inc,
@@ -348,7 +355,7 @@ int rds_message_inc_copy_to_user(struct rds_incoming *inc,
        iov = first_iov;
        iov_off = 0;
-        sg = rm->m_sg;
+        sg = rm->data.op_sg;
        vec_off = 0;
        copied = 0;
@@ -394,15 +401,14 @@ int rds_message_inc_copy_to_user(struct rds_incoming *inc,
 */
 void rds_message_wait(struct rds_message *rm)
 {
-        wait_event(rds_message_flush_waitq,
+        wait_event_interruptible(rm->m_flush_wait,
                        !test_bit(RDS_MSG_MAPPED, &rm->m_flags));
 }
 void rds_message_unmapped(struct rds_message *rm)
 {
        clear_bit(RDS_MSG_MAPPED, &rm->m_flags);
-        if (waitqueue_active(&rds_message_flush_waitq))
+        wake_up_interruptible(&rm->m_flush_wait);
-                wake_up(&rds_message_flush_waitq);
 }
 EXPORT_SYMBOL_GPL(rds_message_unmapped);
diff --git a/net/rds/page.c b/net/rds/page.c
index 595a952d4b17..5e44f5ae7898 100644
--- a/net/rds/page.c
+++ b/net/rds/page.c
@@ -116,7 +116,7 @@ int rds_page_remainder_alloc(struct scatterlist *scat, unsigned long bytes,
        /* jump straight to allocation if we're trying for a huge page */
        if (bytes >= PAGE_SIZE) {
                page = alloc_page(gfp);
-                if (page == NULL) {
+                if (!page) {
                        ret = -ENOMEM;
                } else {
                        sg_set_page(scat, page, PAGE_SIZE, 0);
@@ -162,7 +162,7 @@ int rds_page_remainder_alloc(struct scatterlist *scat, unsigned long bytes,
                rem = &per_cpu(rds_page_remainders, get_cpu());
                local_irq_save(flags);
-                if (page == NULL) {
+                if (!page) {
                        ret = -ENOMEM;
                        break;
                }
@@ -186,6 +186,7 @@ out:
                 ret ? 0 : scat->length);
        return ret;
 }
+EXPORT_SYMBOL_GPL(rds_page_remainder_alloc);
 static int rds_page_remainder_cpu_notify(struct notifier_block *self,
                                         unsigned long action, void *hcpu)
diff --git a/net/rds/rdma.c b/net/rds/rdma.c
index 75fd13bb631b..1a41debca1ce 100644
--- a/net/rds/rdma.c
+++ b/net/rds/rdma.c
@@ -35,7 +35,7 @@
 #include <linux/rbtree.h>
 #include <linux/dma-mapping.h> /* for DMA_*_DEVICE */
-#include "rdma.h"
+#include "rds.h"
 /*
 * XXX
@@ -130,14 +130,22 @@ void rds_rdma_drop_keys(struct rds_sock *rs)
 {
        struct rds_mr *mr;
        struct rb_node *node;
+        unsigned long flags;
        /* Release any MRs associated with this socket */
+        spin_lock_irqsave(&rs->rs_rdma_lock, flags);
        while ((node = rb_first(&rs->rs_rdma_keys))) {
                mr = container_of(node, struct rds_mr, r_rb_node);
                if (mr->r_trans == rs->rs_transport)
                        mr->r_invalidate = 0;
+                rb_erase(&mr->r_rb_node, &rs->rs_rdma_keys);
+                RB_CLEAR_NODE(&mr->r_rb_node);
+                spin_unlock_irqrestore(&rs->rs_rdma_lock, flags);
+                rds_destroy_mr(mr);
                rds_mr_put(mr);
+                spin_lock_irqsave(&rs->rs_rdma_lock, flags);
        }
+        spin_unlock_irqrestore(&rs->rs_rdma_lock, flags);
        if (rs->rs_transport && rs->rs_transport->flush_mrs)
                rs->rs_transport->flush_mrs();
@@ -181,7 +189,7 @@ static int __rds_rdma_map(struct rds_sock *rs, struct rds_get_mr_args *args,
                goto out;
        }
-        if (rs->rs_transport->get_mr == NULL) {
+        if (!rs->rs_transport->get_mr) {
                ret = -EOPNOTSUPP;
                goto out;
        }
@@ -197,13 +205,13 @@ static int __rds_rdma_map(struct rds_sock *rs, struct rds_get_mr_args *args,
        /* XXX clamp nr_pages to limit the size of this alloc? */
        pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
-        if (pages == NULL) {
+        if (!pages) {
                ret = -ENOMEM;
                goto out;
        }
        mr = kzalloc(sizeof(struct rds_mr), GFP_KERNEL);
-        if (mr == NULL) {
+        if (!mr) {
                ret = -ENOMEM;
                goto out;
        }
@@ -230,13 +238,13 @@ static int __rds_rdma_map(struct rds_sock *rs, struct rds_get_mr_args *args,
         * r/o or r/w. We need to assume r/w, or we'll do a lot of RDMA to
         * the zero page.
         */
-        ret = rds_pin_pages(args->vec.addr & PAGE_MASK, nr_pages, pages, 1);
+        ret = rds_pin_pages(args->vec.addr, nr_pages, pages, 1);
        if (ret < 0)
                goto out;
        nents = ret;
        sg = kcalloc(nents, sizeof(*sg), GFP_KERNEL);
-        if (sg == NULL) {
+        if (!sg) {
                ret = -ENOMEM;
                goto out;
        }
@@ -406,68 +414,127 @@ void rds_rdma_unuse(struct rds_sock *rs, u32 r_key, int force)
        spin_lock_irqsave(&rs->rs_rdma_lock, flags);
        mr = rds_mr_tree_walk(&rs->rs_rdma_keys, r_key, NULL);
-        if (mr && (mr->r_use_once || force)) {
+        if (!mr) {
+                printk(KERN_ERR "rds: trying to unuse MR with unknown r_key %u!\n", r_key);
+                spin_unlock_irqrestore(&rs->rs_rdma_lock, flags);
+                return;
+        }
+        if (mr->r_use_once || force) {
                rb_erase(&mr->r_rb_node, &rs->rs_rdma_keys);
                RB_CLEAR_NODE(&mr->r_rb_node);
                zot_me = 1;
-        } else if (mr)
+        }
-                atomic_inc(&mr->r_refcount);
        spin_unlock_irqrestore(&rs->rs_rdma_lock, flags);
        /* May have to issue a dma_sync on this memory region.
         * Note we could avoid this if the operation was a RDMA READ,
         * but at this point we can't tell. */
-        if (mr != NULL) {
+        if (mr->r_trans->sync_mr)
-                if (mr->r_trans->sync_mr)
+                mr->r_trans->sync_mr(mr->r_trans_private, DMA_FROM_DEVICE);
-                        mr->r_trans->sync_mr(mr->r_trans_private, DMA_FROM_DEVICE);
+        /* If the MR was marked as invalidate, this will
-                /* If the MR was marked as invalidate, this will
+         * trigger an async flush. */
-                 * trigger an async flush. */
+        if (zot_me)
-                if (zot_me)
+                rds_destroy_mr(mr);
-                        rds_destroy_mr(mr);
+        rds_mr_put(mr);
-                rds_mr_put(mr);
-        }
 }
-void rds_rdma_free_op(struct rds_rdma_op *ro)
+void rds_rdma_free_op(struct rm_rdma_op *ro)
 {
        unsigned int i;
-        for (i = 0; i < ro->r_nents; i++) {
+        for (i = 0; i < ro->op_nents; i++) {
-                struct page *page = sg_page(&ro->r_sg[i]);
+                struct page *page = sg_page(&ro->op_sg[i]);
                /* Mark page dirty if it was possibly modified, which
                 * is the case for a RDMA_READ which copies from remote
                 * to local memory */
-                if (!ro->r_write) {
+                if (!ro->op_write) {
-                        BUG_ON(in_interrupt());
+                        BUG_ON(irqs_disabled());
                        set_page_dirty(page);
                }
                put_page(page);
        }
-        kfree(ro->r_notifier);
+        kfree(ro->op_notifier);
-        kfree(ro);
+        ro->op_notifier = NULL;
+        ro->op_active = 0;
+}
+void rds_atomic_free_op(struct rm_atomic_op *ao)
+{
+        struct page *page = sg_page(ao->op_sg);
+        /* Mark page dirty if it was possibly modified, which
+         * is the case for a RDMA_READ which copies from remote
+         * to local memory */
+        set_page_dirty(page);
+        put_page(page);
+        kfree(ao->op_notifier);
+        ao->op_notifier = NULL;
+        ao->op_active = 0;
 }
 /*
- * args is a pointer to an in-kernel copy in the sendmsg cmsg.
+ * Count the number of pages needed to describe an incoming iovec.
 */
-static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
+static int rds_rdma_pages(struct rds_rdma_args *args)
-                                            struct rds_rdma_args *args)
 {
        struct rds_iovec vec;
-        struct rds_rdma_op *op = NULL;
+        struct rds_iovec __user *local_vec;
+        unsigned int tot_pages = 0;
        unsigned int nr_pages;
-        unsigned int max_pages;
+        unsigned int i;
+        local_vec = (struct rds_iovec __user *)(unsigned long) args->local_vec_addr;
+        /* figure out the number of pages in the vector */
+        for (i = 0; i < args->nr_local; i++) {
+                if (copy_from_user(&vec, &local_vec[i],
+                                   sizeof(struct rds_iovec)))
+                        return -EFAULT;
+                nr_pages = rds_pages_in_vec(&vec);
+                if (nr_pages == 0)
+                        return -EINVAL;
+                tot_pages += nr_pages;
+        }
+        return tot_pages;
+}
+int rds_rdma_extra_size(struct rds_rdma_args *args)
+{
+        return rds_rdma_pages(args) * sizeof(struct scatterlist);
+}
+/*
+ * The application asks for a RDMA transfer.
+ * Extract all arguments and set up the rdma_op
+ */
+int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
+                          struct cmsghdr *cmsg)
+{
+        struct rds_rdma_args *args;
+        struct rds_iovec vec;
+        struct rm_rdma_op *op = &rm->rdma;
+        int nr_pages;
        unsigned int nr_bytes;
        struct page **pages = NULL;
        struct rds_iovec __user *local_vec;
-        struct scatterlist *sg;
        unsigned int nr;
        unsigned int i, j;
-        int ret;
+        int ret = 0;
+        if (cmsg->cmsg_len < CMSG_LEN(sizeof(struct rds_rdma_args))
+            || rm->rdma.op_active)
+                return -EINVAL;
+        args = CMSG_DATA(cmsg);
        if (rs->rs_bound_addr == 0) {
                ret = -ENOTCONN; /* XXX not a great errno */
@@ -479,61 +546,38 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
                goto out;
        }
-        nr_pages = 0;
+        nr_pages = rds_rdma_pages(args);
-        max_pages = 0;
+        if (nr_pages < 0)
-        local_vec = (struct rds_iovec __user *)(unsigned long) args->local_vec_addr;
-        /* figure out the number of pages in the vector */
-        for (i = 0; i < args->nr_local; i++) {
-                if (copy_from_user(&vec, &local_vec[i],
-                                   sizeof(struct rds_iovec))) {
-                        ret = -EFAULT;
-                        goto out;
-                }
-                nr = rds_pages_in_vec(&vec);
-                if (nr == 0) {
-                        ret = -EINVAL;
-                        goto out;
-                }
-                max_pages = max(nr, max_pages);
-                nr_pages += nr;
-        }
-        pages = kcalloc(max_pages, sizeof(struct page *), GFP_KERNEL);
-        if (pages == NULL) {
-                ret = -ENOMEM;
                goto out;
-        }
-        op = kzalloc(offsetof(struct rds_rdma_op, r_sg[nr_pages]), GFP_KERNEL);
+        pages = kcalloc(nr_pages, sizeof(struct page *), GFP_KERNEL);
-        if (op == NULL) {
+        if (!pages) {
                ret = -ENOMEM;
                goto out;
        }
-        op->r_write = !!(args->flags & RDS_RDMA_READWRITE);
+        op->op_write = !!(args->flags & RDS_RDMA_READWRITE);
-        op->r_fence = !!(args->flags & RDS_RDMA_FENCE);
+        op->op_fence = !!(args->flags & RDS_RDMA_FENCE);
-        op->r_notify = !!(args->flags & RDS_RDMA_NOTIFY_ME);
+        op->op_notify = !!(args->flags & RDS_RDMA_NOTIFY_ME);
-        op->r_recverr = rs->rs_recverr;
+        op->op_silent = !!(args->flags & RDS_RDMA_SILENT);
+        op->op_active = 1;
+        op->op_recverr = rs->rs_recverr;
        WARN_ON(!nr_pages);
-        sg_init_table(op->r_sg, nr_pages);
+        op->op_sg = rds_message_alloc_sgs(rm, nr_pages);
-        if (op->r_notify || op->r_recverr) {
+        if (op->op_notify || op->op_recverr) {
                /* We allocate an uninitialized notifier here, because
                 * we don't want to do that in the completion handler. We
                 * would have to use GFP_ATOMIC there, and don't want to deal
                 * with failed allocations.
                 */
-                op->r_notifier = kmalloc(sizeof(struct rds_notifier), GFP_KERNEL);
+                op->op_notifier = kmalloc(sizeof(struct rds_notifier), GFP_KERNEL);
-                if (!op->r_notifier) {
+                if (!op->op_notifier) {
                        ret = -ENOMEM;
                        goto out;
                }
-                op->r_notifier->n_user_token = args->user_token;
+                op->op_notifier->n_user_token = args->user_token;
-                op->r_notifier->n_status = RDS_RDMA_SUCCESS;
+                op->op_notifier->n_status = RDS_RDMA_SUCCESS;
        }
        /* The cookie contains the R_Key of the remote memory region, and
@@ -543,15 +587,17 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
         * destination address (which is really an offset into the MR)
         * FIXME: We may want to move this into ib_rdma.c
         */
-        op->r_key = rds_rdma_cookie_key(args->cookie);
+        op->op_rkey = rds_rdma_cookie_key(args->cookie);
-        op->r_remote_addr = args->remote_vec.addr + rds_rdma_cookie_offset(args->cookie);
+        op->op_remote_addr = args->remote_vec.addr + rds_rdma_cookie_offset(args->cookie);
        nr_bytes = 0;
        rdsdebug("RDS: rdma prepare nr_local %llu rva %llx rkey %x\n",
               (unsigned long long)args->nr_local,
               (unsigned long long)args->remote_vec.addr,
-               op->r_key);
+               op->op_rkey);
+        local_vec = (struct rds_iovec __user *)(unsigned long) args->local_vec_addr;
        for (i = 0; i < args->nr_local; i++) {
                if (copy_from_user(&vec, &local_vec[i],
@@ -569,15 +615,10 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
                rs->rs_user_addr = vec.addr;
                rs->rs_user_bytes = vec.bytes;
-                /* did the user change the vec under us? */
-                if (nr > max_pages || op->r_nents + nr > nr_pages) {
-                        ret = -EINVAL;
-                        goto out;
-                }
                /* If it's a WRITE operation, we want to pin the pages for reading.
                 * If it's a READ operation, we need to pin the pages for writing.
                 */
-                ret = rds_pin_pages(vec.addr & PAGE_MASK, nr, pages, !op->r_write);
+                ret = rds_pin_pages(vec.addr, nr, pages, !op->op_write);
                if (ret < 0)
                        goto out;
@@ -588,8 +629,9 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
                for (j = 0; j < nr; j++) {
                        unsigned int offset = vec.addr & ~PAGE_MASK;
+                        struct scatterlist *sg;
-                        sg = &op->r_sg[op->r_nents + j];
+                        sg = &op->op_sg[op->op_nents + j];
                        sg_set_page(sg, pages[j],
                                        min_t(unsigned int, vec.bytes, PAGE_SIZE - offset),
                                        offset);
@@ -601,10 +643,9 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
                        vec.bytes -= sg->length;
                }
-                op->r_nents += nr;
+                op->op_nents += nr;
        }
        if (nr_bytes > args->remote_vec.bytes) {
                rdsdebug("RDS nr_bytes %u remote_bytes %u do not match\n",
                                nr_bytes,
@@ -612,38 +653,17 @@ static struct rds_rdma_op *rds_rdma_prepare(struct rds_sock *rs,
                ret = -EINVAL;
                goto out;
        }
-        op->r_bytes = nr_bytes;
+        op->op_bytes = nr_bytes;
        ret = 0;
 out:
        kfree(pages);
-        if (ret) {
+        if (ret)
-                if (op)
+                rds_rdma_free_op(op);
-                        rds_rdma_free_op(op);
-                op = ERR_PTR(ret);
-        }
-        return op;
-}
-/*
- * The application asks for a RDMA transfer.
- * Extract all arguments and set up the rdma_op
- */
-int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-                          struct cmsghdr *cmsg)
-{
-        struct rds_rdma_op *op;
-        if (cmsg->cmsg_len < CMSG_LEN(sizeof(struct rds_rdma_args)) ||
-            rm->m_rdma_op != NULL)
-                return -EINVAL;
-        op = rds_rdma_prepare(rs, CMSG_DATA(cmsg));
-        if (IS_ERR(op))
-                return PTR_ERR(op);
        rds_stats_inc(s_send_rdma);
-        rm->m_rdma_op = op;
-        return 0;
+        return ret;
 }
 /*
@@ -673,7 +693,7 @@ int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
        spin_lock_irqsave(&rs->rs_rdma_lock, flags);
        mr = rds_mr_tree_walk(&rs->rs_rdma_keys, r_key, NULL);
-        if (mr == NULL)
+        if (!mr)
                err = -EINVAL;  /* invalid r_key */
        else
                atomic_inc(&mr->r_refcount);
@@ -681,7 +701,7 @@ int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
        if (mr) {
                mr->r_trans->sync_mr(mr->r_trans_private, DMA_TO_DEVICE);
-                rm->m_rdma_mr = mr;
+                rm->rdma.op_rdma_mr = mr;
        }
        return err;
 }
@@ -699,5 +719,98 @@ int rds_cmsg_rdma_map(struct rds_sock *rs, struct rds_message *rm,
            rm->m_rdma_cookie != 0)
                return -EINVAL;
-        return __rds_rdma_map(rs, CMSG_DATA(cmsg), &rm->m_rdma_cookie, &rm->m_rdma_mr);
+        return __rds_rdma_map(rs, CMSG_DATA(cmsg), &rm->m_rdma_cookie, &rm->rdma.op_rdma_mr);
+}
+/*
+ * Fill in rds_message for an atomic request.
+ */
+int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
+                    struct cmsghdr *cmsg)
+{
+        struct page *page = NULL;
+        struct rds_atomic_args *args;
+        int ret = 0;
+        if (cmsg->cmsg_len < CMSG_LEN(sizeof(struct rds_atomic_args))
+         || rm->atomic.op_active)
+                return -EINVAL;
+        args = CMSG_DATA(cmsg);
+        /* Nonmasked & masked cmsg ops converted to masked hw ops */
+        switch (cmsg->cmsg_type) {
+        case RDS_CMSG_ATOMIC_FADD:
+                rm->atomic.op_type = RDS_ATOMIC_TYPE_FADD;
+                rm->atomic.op_m_fadd.add = args->fadd.add;
+                rm->atomic.op_m_fadd.nocarry_mask = 0;
+                break;
+        case RDS_CMSG_MASKED_ATOMIC_FADD:
+                rm->atomic.op_type = RDS_ATOMIC_TYPE_FADD;
+                rm->atomic.op_m_fadd.add = args->m_fadd.add;
+                rm->atomic.op_m_fadd.nocarry_mask = args->m_fadd.nocarry_mask;
+                break;
+        case RDS_CMSG_ATOMIC_CSWP:
+                rm->atomic.op_type = RDS_ATOMIC_TYPE_CSWP;
+                rm->atomic.op_m_cswp.compare = args->cswp.compare;
+                rm->atomic.op_m_cswp.swap = args->cswp.swap;
+                rm->atomic.op_m_cswp.compare_mask = ~0;
+                rm->atomic.op_m_cswp.swap_mask = ~0;
+                break;
+        case RDS_CMSG_MASKED_ATOMIC_CSWP:
+                rm->atomic.op_type = RDS_ATOMIC_TYPE_CSWP;
+                rm->atomic.op_m_cswp.compare = args->m_cswp.compare;
+                rm->atomic.op_m_cswp.swap = args->m_cswp.swap;
+                rm->atomic.op_m_cswp.compare_mask = args->m_cswp.compare_mask;
+                rm->atomic.op_m_cswp.swap_mask = args->m_cswp.swap_mask;
+                break;
+        default:
+                BUG(); /* should never happen */
+        }
+        rm->atomic.op_notify = !!(args->flags & RDS_RDMA_NOTIFY_ME);
+        rm->atomic.op_silent = !!(args->flags & RDS_RDMA_SILENT);
+        rm->atomic.op_active = 1;
+        rm->atomic.op_recverr = rs->rs_recverr;
+        rm->atomic.op_sg = rds_message_alloc_sgs(rm, 1);
+        /* verify 8 byte-aligned */
+        if (args->local_addr & 0x7) {
+                ret = -EFAULT;
+                goto err;
+        }
+        ret = rds_pin_pages(args->local_addr, 1, &page, 1);
+        if (ret != 1)
+                goto err;
+        ret = 0;
+        sg_set_page(rm->atomic.op_sg, page, 8, offset_in_page(args->local_addr));
+        if (rm->atomic.op_notify || rm->atomic.op_recverr) {
+                /* We allocate an uninitialized notifier here, because
+                 * we don't want to do that in the completion handler. We
+                 * would have to use GFP_ATOMIC there, and don't want to deal
+                 * with failed allocations.
+                 */
+                rm->atomic.op_notifier = kmalloc(sizeof(*rm->atomic.op_notifier), GFP_KERNEL);
+                if (!rm->atomic.op_notifier) {
+                        ret = -ENOMEM;
+                        goto err;
+                }
+                rm->atomic.op_notifier->n_user_token = args->user_token;
+                rm->atomic.op_notifier->n_status = RDS_RDMA_SUCCESS;
+        }
+        rm->atomic.op_rkey = rds_rdma_cookie_key(args->cookie);
+        rm->atomic.op_remote_addr = args->remote_addr + rds_rdma_cookie_offset(args->cookie);
+        return ret;
+err:
+        if (page)
+                put_page(page);
+        kfree(rm->atomic.op_notifier);
+        return ret;
 }
diff --git a/net/rds/rdma.h b/net/rds/rdma.h
deleted file mode 100644
index 909c39835a5d..000000000000
--- a/net/rds/rdma.h
+++ /dev/null
@@ -1,85 +0,0 @@
-#ifndef _RDS_RDMA_H
-#define _RDS_RDMA_H
-#include <linux/rbtree.h>
-#include <linux/spinlock.h>
-#include <linux/scatterlist.h>
-#include "rds.h"
-struct rds_mr {
-        struct rb_node          r_rb_node;
-        atomic_t                r_refcount;
-        u32                     r_key;
-        /* A copy of the creation flags */
-        unsigned int            r_use_once:1;
-        unsigned int            r_invalidate:1;
-        unsigned int            r_write:1;
-        /* This is for RDS_MR_DEAD.
-         * It would be nice & consistent to make this part of the above
-         * bit field here, but we need to use test_and_set_bit.
-         */
-        unsigned long           r_state;
-        struct rds_sock         *r_sock; /* back pointer to the socket that owns us */
-        struct rds_transport    *r_trans;
-        void                    *r_trans_private;
-};
-/* Flags for mr->r_state */
-#define RDS_MR_DEAD             0
-struct rds_rdma_op {
-        u32                     r_key;
-        u64                     r_remote_addr;
-        unsigned int            r_write:1;
-        unsigned int            r_fence:1;
-        unsigned int            r_notify:1;
-        unsigned int            r_recverr:1;
-        unsigned int            r_mapped:1;
-        struct rds_notifier     *r_notifier;
-        unsigned int            r_bytes;
-        unsigned int            r_nents;
-        unsigned int            r_count;
-        struct scatterlist      r_sg[0];
-};
-static inline rds_rdma_cookie_t rds_rdma_make_cookie(u32 r_key, u32 offset)
-{
-        return r_key | (((u64) offset) << 32);
-}
-static inline u32 rds_rdma_cookie_key(rds_rdma_cookie_t cookie)
-{
-        return cookie;
-}
-static inline u32 rds_rdma_cookie_offset(rds_rdma_cookie_t cookie)
-{
-        return cookie >> 32;
-}
-int rds_get_mr(struct rds_sock *rs, char __user *optval, int optlen);
-int rds_get_mr_for_dest(struct rds_sock *rs, char __user *optval, int optlen);
-int rds_free_mr(struct rds_sock *rs, char __user *optval, int optlen);
-void rds_rdma_drop_keys(struct rds_sock *rs);
-int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-                          struct cmsghdr *cmsg);
-int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
-                          struct cmsghdr *cmsg);
-int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
-                          struct cmsghdr *cmsg);
-int rds_cmsg_rdma_map(struct rds_sock *rs, struct rds_message *rm,
-                          struct cmsghdr *cmsg);
-void rds_rdma_free_op(struct rds_rdma_op *ro);
-void rds_rdma_send_complete(struct rds_message *rm, int);
-extern void __rds_put_mr_final(struct rds_mr *mr);
-static inline void rds_mr_put(struct rds_mr *mr)
-{
-        if (atomic_dec_and_test(&mr->r_refcount))
-                __rds_put_mr_final(mr);
-}
-#endif
diff --git a/net/rds/rdma_transport.c b/net/rds/rdma_transport.c
index e599ba2f950d..e6ed10aee190 100644
--- a/net/rds/rdma_transport.c
+++ b/net/rds/rdma_transport.c
@@ -36,6 +36,34 @@
 static struct rdma_cm_id *rds_rdma_listen_id;
+static char *rds_cm_event_strings[] = {
+#define RDS_CM_EVENT_STRING(foo) \
+                [RDMA_CM_EVENT_##foo] = __stringify(RDMA_CM_EVENT_##foo)
+        RDS_CM_EVENT_STRING(ADDR_RESOLVED),
+        RDS_CM_EVENT_STRING(ADDR_ERROR),
+        RDS_CM_EVENT_STRING(ROUTE_RESOLVED),
+        RDS_CM_EVENT_STRING(ROUTE_ERROR),
+        RDS_CM_EVENT_STRING(CONNECT_REQUEST),
+        RDS_CM_EVENT_STRING(CONNECT_RESPONSE),
+        RDS_CM_EVENT_STRING(CONNECT_ERROR),
+        RDS_CM_EVENT_STRING(UNREACHABLE),
+        RDS_CM_EVENT_STRING(REJECTED),
+        RDS_CM_EVENT_STRING(ESTABLISHED),
+        RDS_CM_EVENT_STRING(DISCONNECTED),
+        RDS_CM_EVENT_STRING(DEVICE_REMOVAL),
+        RDS_CM_EVENT_STRING(MULTICAST_JOIN),
+        RDS_CM_EVENT_STRING(MULTICAST_ERROR),
+        RDS_CM_EVENT_STRING(ADDR_CHANGE),
+        RDS_CM_EVENT_STRING(TIMEWAIT_EXIT),
+#undef RDS_CM_EVENT_STRING
+};
+static char *rds_cm_event_str(enum rdma_cm_event_type type)
+{
+        return rds_str_array(rds_cm_event_strings,
+                             ARRAY_SIZE(rds_cm_event_strings), type);
+};
 int rds_rdma_cm_event_handler(struct rdma_cm_id *cm_id,
                              struct rdma_cm_event *event)
 {
@@ -44,8 +72,8 @@ int rds_rdma_cm_event_handler(struct rdma_cm_id *cm_id,
        struct rds_transport *trans;
        int ret = 0;
-        rdsdebug("conn %p id %p handling event %u\n", conn, cm_id,
+        rdsdebug("conn %p id %p handling event %u (%s)\n", conn, cm_id,
-                 event->event);
+                 event->event, rds_cm_event_str(event->event));
        if (cm_id->device->node_type == RDMA_NODE_RNIC)
                trans = &rds_iw_transport;
@@ -109,7 +137,8 @@ int rds_rdma_cm_event_handler(struct rdma_cm_id *cm_id,
        default:
                /* things like device disconnect? */
-                printk(KERN_ERR "RDS: unknown event %u!\n", event->event);
+                printk(KERN_ERR "RDS: unknown event %u (%s)!\n",
+                       event->event, rds_cm_event_str(event->event));
                break;
        }
@@ -117,12 +146,13 @@ out:
        if (conn)
                mutex_unlock(&conn->c_cm_lock);
-        rdsdebug("id %p event %u handling ret %d\n", cm_id, event->event, ret);
+        rdsdebug("id %p event %u (%s) handling ret %d\n", cm_id, event->event,
+                 rds_cm_event_str(event->event), ret);
        return ret;
 }
-static int __init rds_rdma_listen_init(void)
+static int rds_rdma_listen_init(void)
 {
        struct sockaddr_in sin;
        struct rdma_cm_id *cm_id;
@@ -177,7 +207,7 @@ static void rds_rdma_listen_stop(void)
        }
 }
-int __init rds_rdma_init(void)
+int rds_rdma_init(void)
 {
        int ret;
diff --git a/net/rds/rds.h b/net/rds/rds.h
index c224b5bb3ba9..8103dcf8b976 100644
--- a/net/rds/rds.h
+++ b/net/rds/rds.h
@@ -80,6 +80,7 @@ enum {
 /* Bits for c_flags */
 #define RDS_LL_SEND_FULL        0
 #define RDS_RECONNECT_PENDING   1
+#define RDS_IN_XMIT             2
 struct rds_connection {
        struct hlist_node       c_hash_node;
@@ -91,12 +92,13 @@ struct rds_connection {
        struct rds_cong_map     *c_lcong;
        struct rds_cong_map     *c_fcong;
-        struct mutex            c_send_lock;    /* protect send ring */
        struct rds_message      *c_xmit_rm;
        unsigned long           c_xmit_sg;
        unsigned int            c_xmit_hdr_off;
        unsigned int            c_xmit_data_off;
+        unsigned int            c_xmit_atomic_sent;
        unsigned int            c_xmit_rdma_sent;
+        unsigned int            c_xmit_data_sent;
        spinlock_t              c_lock;         /* protect msg queues */
        u64                     c_next_tx_seq;
@@ -116,11 +118,10 @@ struct rds_connection {
        struct delayed_work     c_conn_w;
        struct work_struct      c_down_w;
        struct mutex            c_cm_lock;      /* protect conn state & cm */
+        wait_queue_head_t       c_waitq;
        struct list_head        c_map_item;
        unsigned long           c_map_queued;
-        unsigned long           c_map_offset;
-        unsigned long           c_map_bytes;
        unsigned int            c_unacked_packets;
        unsigned int            c_unacked_bytes;
@@ -206,6 +207,48 @@ struct rds_incoming {
        rds_rdma_cookie_t       i_rdma_cookie;
 };
+struct rds_mr {
+        struct rb_node          r_rb_node;
+        atomic_t                r_refcount;
+        u32                     r_key;
+        /* A copy of the creation flags */
+        unsigned int            r_use_once:1;
+        unsigned int            r_invalidate:1;
+        unsigned int            r_write:1;
+        /* This is for RDS_MR_DEAD.
+         * It would be nice & consistent to make this part of the above
+         * bit field here, but we need to use test_and_set_bit.
+         */
+        unsigned long           r_state;
+        struct rds_sock         *r_sock; /* back pointer to the socket that owns us */
+        struct rds_transport    *r_trans;
+        void                    *r_trans_private;
+};
+/* Flags for mr->r_state */
+#define RDS_MR_DEAD             0
+static inline rds_rdma_cookie_t rds_rdma_make_cookie(u32 r_key, u32 offset)
+{
+        return r_key | (((u64) offset) << 32);
+}
+static inline u32 rds_rdma_cookie_key(rds_rdma_cookie_t cookie)
+{
+        return cookie;
+}
+static inline u32 rds_rdma_cookie_offset(rds_rdma_cookie_t cookie)
+{
+        return cookie >> 32;
+}
+/* atomic operation types */
+#define RDS_ATOMIC_TYPE_CSWP            0
+#define RDS_ATOMIC_TYPE_FADD            1
 /*
 * m_sock_item and m_conn_item are on lists that are serialized under
 * conn->c_lock.  m_sock_item has additional meaning in that once it is empty
@@ -258,13 +301,71 @@ struct rds_message {
         *   -> rs->rs_lock
         */
        spinlock_t              m_rs_lock;
+        wait_queue_head_t       m_flush_wait;
        struct rds_sock         *m_rs;
-        struct rds_rdma_op      *m_rdma_op;
+        /* cookie to send to remote, in rds header */
        rds_rdma_cookie_t       m_rdma_cookie;
-        struct rds_mr           *m_rdma_mr;
-        unsigned int            m_nents;
+        unsigned int            m_used_sgs;
-        unsigned int            m_count;
+        unsigned int            m_total_sgs;
-        struct scatterlist      m_sg[0];
+        void                    *m_final_op;
+        struct {
+                struct rm_atomic_op {
+                        int                     op_type;
+                        union {
+                                struct {
+                                        uint64_t        compare;
+                                        uint64_t        swap;
+                                        uint64_t        compare_mask;
+                                        uint64_t        swap_mask;
+                                } op_m_cswp;
+                                struct {
+                                        uint64_t        add;
+                                        uint64_t        nocarry_mask;
+                                } op_m_fadd;
+                        };
+                        u32                     op_rkey;
+                        u64                     op_remote_addr;
+                        unsigned int            op_notify:1;
+                        unsigned int            op_recverr:1;
+                        unsigned int            op_mapped:1;
+                        unsigned int            op_silent:1;
+                        unsigned int            op_active:1;
+                        struct scatterlist      *op_sg;
+                        struct rds_notifier     *op_notifier;
+                        struct rds_mr           *op_rdma_mr;
+                } atomic;
+                struct rm_rdma_op {
+                        u32                     op_rkey;
+                        u64                     op_remote_addr;
+                        unsigned int            op_write:1;
+                        unsigned int            op_fence:1;
+                        unsigned int            op_notify:1;
+                        unsigned int            op_recverr:1;
+                        unsigned int            op_mapped:1;
+                        unsigned int            op_silent:1;
+                        unsigned int            op_active:1;
+                        unsigned int            op_bytes;
+                        unsigned int            op_nents;
+                        unsigned int            op_count;
+                        struct scatterlist      *op_sg;
+                        struct rds_notifier     *op_notifier;
+                        struct rds_mr           *op_rdma_mr;
+                } rdma;
+                struct rm_data_op {
+                        unsigned int            op_active:1;
+                        unsigned int            op_nents;
+                        unsigned int            op_count;
+                        struct scatterlist      *op_sg;
+                } data;
+        };
 };
 /*
@@ -305,10 +406,6 @@ struct rds_notifier {
 *                 transport is responsible for other serialization, including
 *                 rds_recv_incoming().  This is called in process context but
 *                 should try hard not to block.
- *
- * @xmit_cong_map: This asks the transport to send the local bitmap down the
- *                 given connection.  XXX get a better story about the bitmap
- *                 flag and header.
 */
 #define RDS_TRANS_IB    0
@@ -332,13 +429,11 @@ struct rds_transport {
        void (*xmit_complete)(struct rds_connection *conn);
        int (*xmit)(struct rds_connection *conn, struct rds_message *rm,
                    unsigned int hdr_off, unsigned int sg, unsigned int off);
-        int (*xmit_cong_map)(struct rds_connection *conn,
+        int (*xmit_rdma)(struct rds_connection *conn, struct rm_rdma_op *op);
-                             struct rds_cong_map *map, unsigned long offset);
+        int (*xmit_atomic)(struct rds_connection *conn, struct rm_atomic_op *op);
-        int (*xmit_rdma)(struct rds_connection *conn, struct rds_rdma_op *op);
        int (*recv)(struct rds_connection *conn);
        int (*inc_copy_to_user)(struct rds_incoming *inc, struct iovec *iov,
                                size_t size);
-        void (*inc_purge)(struct rds_incoming *inc);
        void (*inc_free)(struct rds_incoming *inc);
        int (*cm_handle_connect)(struct rdma_cm_id *cm_id,
@@ -367,17 +462,11 @@ struct rds_sock {
         * bound_addr used for both incoming and outgoing, no INADDR_ANY
         * support.
         */
-        struct rb_node          rs_bound_node;
+        struct hlist_node       rs_bound_node;
        __be32                  rs_bound_addr;
        __be32                  rs_conn_addr;
        __be16                  rs_bound_port;
        __be16                  rs_conn_port;
-        /*
-         * This is only used to communicate the transport between bind and
-         * initiating connections.  All other trans use is referenced through
-         * the connection.
-         */
        struct rds_transport    *rs_transport;
        /*
@@ -466,8 +555,8 @@ struct rds_statistics {
        uint64_t        s_recv_ping;
        uint64_t        s_send_queue_empty;
        uint64_t        s_send_queue_full;
-        uint64_t        s_send_sem_contention;
+        uint64_t        s_send_lock_contention;
-        uint64_t        s_send_sem_queue_raced;
+        uint64_t        s_send_lock_queue_raced;
        uint64_t        s_send_immediate_retry;
        uint64_t        s_send_delayed_retry;
        uint64_t        s_send_drop_acked;
@@ -487,6 +576,7 @@ struct rds_statistics {
 };
 /* af_rds.c */
+char *rds_str_array(char **array, size_t elements, size_t index);
 void rds_sock_addref(struct rds_sock *rs);
 void rds_sock_put(struct rds_sock *rs);
 void rds_wake_sk_sleep(struct rds_sock *rs);
@@ -521,15 +611,17 @@ void rds_cong_exit(void);
 struct rds_message *rds_cong_update_alloc(struct rds_connection *conn);
 /* conn.c */
-int __init rds_conn_init(void);
+int rds_conn_init(void);
 void rds_conn_exit(void);
 struct rds_connection *rds_conn_create(__be32 laddr, __be32 faddr,
                                       struct rds_transport *trans, gfp_t gfp);
 struct rds_connection *rds_conn_create_outgoing(__be32 laddr, __be32 faddr,
                               struct rds_transport *trans, gfp_t gfp);
+void rds_conn_shutdown(struct rds_connection *conn);
 void rds_conn_destroy(struct rds_connection *conn);
 void rds_conn_reset(struct rds_connection *conn);
 void rds_conn_drop(struct rds_connection *conn);
+void rds_conn_connect_if_down(struct rds_connection *conn);
 void rds_for_each_conn_info(struct socket *sock, unsigned int len,
                          struct rds_info_iterator *iter,
                          struct rds_info_lengths *lens,
@@ -566,7 +658,8 @@ rds_conn_connecting(struct rds_connection *conn)
 /* message.c */
 struct rds_message *rds_message_alloc(unsigned int nents, gfp_t gfp);
-struct rds_message *rds_message_copy_from_user(struct iovec *first_iov,
+struct scatterlist *rds_message_alloc_sgs(struct rds_message *rm, int nents);
+int rds_message_copy_from_user(struct rds_message *rm, struct iovec *first_iov,
                                               size_t total_len);
 struct rds_message *rds_message_map_pages(unsigned long *page_addrs, unsigned int total_len);
 void rds_message_populate_header(struct rds_header *hdr, __be16 sport,
@@ -580,7 +673,6 @@ int rds_message_get_version_extension(struct rds_header *hdr, unsigned int *vers
 int rds_message_add_rdma_dest_extension(struct rds_header *hdr, u32 r_key, u32 offset);
 int rds_message_inc_copy_to_user(struct rds_incoming *inc,
                                 struct iovec *first_iov, size_t size);
-void rds_message_inc_purge(struct rds_incoming *inc);
 void rds_message_inc_free(struct rds_incoming *inc);
 void rds_message_addref(struct rds_message *rm);
 void rds_message_put(struct rds_message *rm);
@@ -636,14 +728,39 @@ void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in *dest);
 typedef int (*is_acked_func)(struct rds_message *rm, uint64_t ack);
 void rds_send_drop_acked(struct rds_connection *conn, u64 ack,
                         is_acked_func is_acked);
-int rds_send_acked_before(struct rds_connection *conn, u64 seq);
 void rds_send_remove_from_sock(struct list_head *messages, int status);
 int rds_send_pong(struct rds_connection *conn, __be16 dport);
 struct rds_message *rds_send_get_message(struct rds_connection *,
-                                         struct rds_rdma_op *);
+                                         struct rm_rdma_op *);
 /* rdma.c */
 void rds_rdma_unuse(struct rds_sock *rs, u32 r_key, int force);
+int rds_get_mr(struct rds_sock *rs, char __user *optval, int optlen);
+int rds_get_mr_for_dest(struct rds_sock *rs, char __user *optval, int optlen);
+int rds_free_mr(struct rds_sock *rs, char __user *optval, int optlen);
+void rds_rdma_drop_keys(struct rds_sock *rs);
+int rds_rdma_extra_size(struct rds_rdma_args *args);
+int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
+                          struct cmsghdr *cmsg);
+int rds_cmsg_rdma_dest(struct rds_sock *rs, struct rds_message *rm,
+                          struct cmsghdr *cmsg);
+int rds_cmsg_rdma_args(struct rds_sock *rs, struct rds_message *rm,
+                          struct cmsghdr *cmsg);
+int rds_cmsg_rdma_map(struct rds_sock *rs, struct rds_message *rm,
+                          struct cmsghdr *cmsg);
+void rds_rdma_free_op(struct rm_rdma_op *ro);
+void rds_atomic_free_op(struct rm_atomic_op *ao);
+void rds_rdma_send_complete(struct rds_message *rm, int wc_status);
+void rds_atomic_send_complete(struct rds_message *rm, int wc_status);
+int rds_cmsg_atomic(struct rds_sock *rs, struct rds_message *rm,
+                    struct cmsghdr *cmsg);
+extern void __rds_put_mr_final(struct rds_mr *mr);
+static inline void rds_mr_put(struct rds_mr *mr)
+{
+        if (atomic_dec_and_test(&mr->r_refcount))
+                __rds_put_mr_final(mr);
+}
 /* stats.c */
 DECLARE_PER_CPU_SHARED_ALIGNED(struct rds_statistics, rds_stats);
@@ -657,14 +774,14 @@ DECLARE_PER_CPU_SHARED_ALIGNED(struct rds_statistics, rds_stats);
        put_cpu();                                      \
 } while (0)
 #define rds_stats_add(member, count) rds_stats_add_which(rds_stats, member, count)
-int __init rds_stats_init(void);
+int rds_stats_init(void);
 void rds_stats_exit(void);
 void rds_stats_info_copy(struct rds_info_iterator *iter,
                         uint64_t *values, const char *const *names,
                         size_t nr);
 /* sysctl.c */
-int __init rds_sysctl_init(void);
+int rds_sysctl_init(void);
 void rds_sysctl_exit(void);
 extern unsigned long rds_sysctl_sndbuf_min;
 extern unsigned long rds_sysctl_sndbuf_default;
@@ -678,9 +795,10 @@ extern unsigned long rds_sysctl_trace_flags;
 extern unsigned int  rds_sysctl_trace_level;
 /* threads.c */
-int __init rds_threads_init(void);
+int rds_threads_init(void);
 void rds_threads_exit(void);
 extern struct workqueue_struct *rds_wq;
+void rds_queue_reconnect(struct rds_connection *conn);
 void rds_connect_worker(struct work_struct *);
 void rds_shutdown_worker(struct work_struct *);
 void rds_send_worker(struct work_struct *);
@@ -691,9 +809,10 @@ void rds_connect_complete(struct rds_connection *conn);
 int rds_trans_register(struct rds_transport *trans);
 void rds_trans_unregister(struct rds_transport *trans);
 struct rds_transport *rds_trans_get_preferred(__be32 addr);
+void rds_trans_put(struct rds_transport *trans);
 unsigned int rds_trans_stats_info_copy(struct rds_info_iterator *iter,
                                       unsigned int avail);
-int __init rds_trans_init(void);
+int rds_trans_init(void);
 void rds_trans_exit(void);
 #endif
diff --git a/net/rds/recv.c b/net/rds/recv.c
index c93588c2d553..68800f02aa30 100644
--- a/net/rds/recv.c
+++ b/net/rds/recv.c
@@ -36,7 +36,6 @@
 #include <linux/in.h>
 #include "rds.h"
-#include "rdma.h"
 void rds_inc_init(struct rds_incoming *inc, struct rds_connection *conn,
                  __be32 saddr)
@@ -210,7 +209,7 @@ void rds_recv_incoming(struct rds_connection *conn, __be32 saddr, __be32 daddr,
        }
        rs = rds_find_bound(daddr, inc->i_hdr.h_dport);
-        if (rs == NULL) {
+        if (!rs) {
                rds_stats_inc(s_recv_drop_no_sock);
                goto out;
        }
@@ -251,7 +250,7 @@ static int rds_next_incoming(struct rds_sock *rs, struct rds_incoming **inc)
 {
        unsigned long flags;
-        if (*inc == NULL) {
+        if (!*inc) {
                read_lock_irqsave(&rs->rs_recv_lock, flags);
                if (!list_empty(&rs->rs_recv_queue)) {
                        *inc = list_entry(rs->rs_recv_queue.next,
@@ -334,10 +333,10 @@ int rds_notify_queue_get(struct rds_sock *rs, struct msghdr *msghdr)
                if (msghdr) {
                        cmsg.user_token = notifier->n_user_token;
-                        cmsg.status  = notifier->n_status;
+                        cmsg.status = notifier->n_status;
                        err = put_cmsg(msghdr, SOL_RDS, RDS_CMSG_RDMA_STATUS,
-                                        sizeof(cmsg), &cmsg);
+                                       sizeof(cmsg), &cmsg);
                        if (err)
                                break;
                }
diff --git a/net/rds/send.c b/net/rds/send.c
index 9c1c6bcaa6c9..9b951a0ab6b7 100644
--- a/net/rds/send.c
+++ b/net/rds/send.c
@@ -37,7 +37,6 @@
 #include <linux/list.h>
 #include "rds.h"
-#include "rdma.h"
 /* When transmitting messages in rds_send_xmit, we need to emerge from
 * time to time and briefly release the CPU. Otherwise the softlock watchdog
@@ -54,7 +53,8 @@ module_param(send_batch_count, int, 0444);
 MODULE_PARM_DESC(send_batch_count, " batch factor when working the send queue");
 /*
- * Reset the send state. Caller must hold c_send_lock when calling here.
+ * Reset the send state.  Callers must ensure that this doesn't race with
+ * rds_send_xmit().
 */
 void rds_send_reset(struct rds_connection *conn)
 {
@@ -62,18 +62,22 @@ void rds_send_reset(struct rds_connection *conn)
        unsigned long flags;
        if (conn->c_xmit_rm) {
+                rm = conn->c_xmit_rm;
+                conn->c_xmit_rm = NULL;
                /* Tell the user the RDMA op is no longer mapped by the
                 * transport. This isn't entirely true (it's flushed out
                 * independently) but as the connection is down, there's
                 * no ongoing RDMA to/from that memory */
-                rds_message_unmapped(conn->c_xmit_rm);
+                rds_message_unmapped(rm);
-                rds_message_put(conn->c_xmit_rm);
+                rds_message_put(rm);
-                conn->c_xmit_rm = NULL;
        }
        conn->c_xmit_sg = 0;
        conn->c_xmit_hdr_off = 0;
        conn->c_xmit_data_off = 0;
+        conn->c_xmit_atomic_sent = 0;
        conn->c_xmit_rdma_sent = 0;
+        conn->c_xmit_data_sent = 0;
        conn->c_map_queued = 0;
@@ -90,6 +94,25 @@ void rds_send_reset(struct rds_connection *conn)
        spin_unlock_irqrestore(&conn->c_lock, flags);
 }
+static int acquire_in_xmit(struct rds_connection *conn)
+{
+        return test_and_set_bit(RDS_IN_XMIT, &conn->c_flags) == 0;
+}
+static void release_in_xmit(struct rds_connection *conn)
+{
+        clear_bit(RDS_IN_XMIT, &conn->c_flags);
+        smp_mb__after_clear_bit();
+        /*
+         * We don't use wait_on_bit()/wake_up_bit() because our waking is in a
+         * hot path and finding waiters is very rare.  We don't want to walk
+         * the system-wide hashed waitqueue buckets in the fast path only to
+         * almost never find waiters.
+         */
+        if (waitqueue_active(&conn->c_waitq))
+                wake_up_all(&conn->c_waitq);
+}
 /*
 * We're making the concious trade-off here to only send one message
 * down the connection at a time.
@@ -109,102 +132,69 @@ int rds_send_xmit(struct rds_connection *conn)
        struct rds_message *rm;
        unsigned long flags;
        unsigned int tmp;
-        unsigned int send_quota = send_batch_count;
        struct scatterlist *sg;
        int ret = 0;
-        int was_empty = 0;
        LIST_HEAD(to_be_dropped);
+restart:
        /*
         * sendmsg calls here after having queued its message on the send
         * queue.  We only have one task feeding the connection at a time.  If
         * another thread is already feeding the queue then we back off.  This
         * avoids blocking the caller and trading per-connection data between
         * caches per message.
-         *
-         * The sem holder will issue a retry if they notice that someone queued
-         * a message after they stopped walking the send queue but before they
-         * dropped the sem.
         */
-        if (!mutex_trylock(&conn->c_send_lock)) {
+        if (!acquire_in_xmit(conn)) {
-                rds_stats_inc(s_send_sem_contention);
+                rds_stats_inc(s_send_lock_contention);
                ret = -ENOMEM;
                goto out;
        }
+        /*
+         * rds_conn_shutdown() sets the conn state and then tests RDS_IN_XMIT,
+         * we do the opposite to avoid races.
+         */
+        if (!rds_conn_up(conn)) {
+                release_in_xmit(conn);
+                ret = 0;
+                goto out;
+        }
        if (conn->c_trans->xmit_prepare)
                conn->c_trans->xmit_prepare(conn);
        /*
         * spin trying to push headers and data down the connection until
-         * the connection doens't make forward progress.
+         * the connection doesn't make forward progress.
         */
-        while (--send_quota) {
+        while (1) {
-                /*
-                 * See if need to send a congestion map update if we're
-                 * between sending messages.  The send_sem protects our sole
-                 * use of c_map_offset and _bytes.
-                 * Note this is used only by transports that define a special
-                 * xmit_cong_map function. For all others, we create allocate
-                 * a cong_map message and treat it just like any other send.
-                 */
-                if (conn->c_map_bytes) {
-                        ret = conn->c_trans->xmit_cong_map(conn, conn->c_lcong,
-                                                conn->c_map_offset);
-                        if (ret <= 0)
-                                break;
-                        conn->c_map_offset += ret;
-                        conn->c_map_bytes -= ret;
-                        if (conn->c_map_bytes)
-                                continue;
-                }
-                /* If we're done sending the current message, clear the
-                 * offset and S/G temporaries.
-                 */
                rm = conn->c_xmit_rm;
-                if (rm != NULL &&
-                    conn->c_xmit_hdr_off == sizeof(struct rds_header) &&
-                    conn->c_xmit_sg == rm->m_nents) {
-                        conn->c_xmit_rm = NULL;
-                        conn->c_xmit_sg = 0;
-                        conn->c_xmit_hdr_off = 0;
-                        conn->c_xmit_data_off = 0;
-                        conn->c_xmit_rdma_sent = 0;
-                        /* Release the reference to the previous message. */
+                /*
-                        rds_message_put(rm);
+                 * If between sending messages, we can send a pending congestion
-                        rm = NULL;
+                 * map update.
-                }
-                /* If we're asked to send a cong map update, do so.
                 */
-                if (rm == NULL && test_and_clear_bit(0, &conn->c_map_queued)) {
+                if (!rm && test_and_clear_bit(0, &conn->c_map_queued)) {
-                        if (conn->c_trans->xmit_cong_map != NULL) {
-                                conn->c_map_offset = 0;
-                                conn->c_map_bytes = sizeof(struct rds_header) +
-                                        RDS_CONG_MAP_BYTES;
-                                continue;
-                        }
                        rm = rds_cong_update_alloc(conn);
                        if (IS_ERR(rm)) {
                                ret = PTR_ERR(rm);
                                break;
                        }
+                        rm->data.op_active = 1;
                        conn->c_xmit_rm = rm;
                }
                /*
-                 * Grab the next message from the send queue, if there is one.
+                 * If not already working on one, grab the next message.
                 *
                 * c_xmit_rm holds a ref while we're sending this message down
                 * the connction.  We can use this ref while holding the
                 * send_sem.. rds_send_reset() is serialized with it.
                 */
-                if (rm == NULL) {
+                if (!rm) {
                        unsigned int len;
                        spin_lock_irqsave(&conn->c_lock, flags);
@@ -224,10 +214,8 @@ int rds_send_xmit(struct rds_connection *conn)
                        spin_unlock_irqrestore(&conn->c_lock, flags);
-                        if (rm == NULL) {
+                        if (!rm)
-                                was_empty = 1;
                                break;
-                        }
                        /* Unfortunately, the way Infiniband deals with
                         * RDMA to a bad MR key is by moving the entire
@@ -236,13 +224,12 @@ int rds_send_xmit(struct rds_connection *conn)
                         * connection.
                         * Therefore, we never retransmit messages with RDMA ops.
                         */
-                        if (rm->m_rdma_op &&
+                        if (rm->rdma.op_active &&
                            test_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags)) {
                                spin_lock_irqsave(&conn->c_lock, flags);
                                if (test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags))
                                        list_move(&rm->m_conn_item, &to_be_dropped);
                                spin_unlock_irqrestore(&conn->c_lock, flags);
-                                rds_message_put(rm);
                                continue;
                        }
@@ -263,23 +250,55 @@ int rds_send_xmit(struct rds_connection *conn)
                        conn->c_xmit_rm = rm;
                }
-                /*
+                /* The transport either sends the whole rdma or none of it */
-                 * Try and send an rdma message.  Let's see if we can
+                if (rm->rdma.op_active && !conn->c_xmit_rdma_sent) {
-                 * keep this simple and require that the transport either
+                        rm->m_final_op = &rm->rdma;
-                 * send the whole rdma or none of it.
+                        ret = conn->c_trans->xmit_rdma(conn, &rm->rdma);
-                 */
-                if (rm->m_rdma_op && !conn->c_xmit_rdma_sent) {
-                        ret = conn->c_trans->xmit_rdma(conn, rm->m_rdma_op);
                        if (ret)
                                break;
                        conn->c_xmit_rdma_sent = 1;
                        /* The transport owns the mapped memory for now.
                         * You can't unmap it while it's on the send queue */
                        set_bit(RDS_MSG_MAPPED, &rm->m_flags);
                }
-                if (conn->c_xmit_hdr_off < sizeof(struct rds_header) ||
+                if (rm->atomic.op_active && !conn->c_xmit_atomic_sent) {
-                    conn->c_xmit_sg < rm->m_nents) {
+                        rm->m_final_op = &rm->atomic;
+                        ret = conn->c_trans->xmit_atomic(conn, &rm->atomic);
+                        if (ret)
+                                break;
+                        conn->c_xmit_atomic_sent = 1;
+                        /* The transport owns the mapped memory for now.
+                         * You can't unmap it while it's on the send queue */
+                        set_bit(RDS_MSG_MAPPED, &rm->m_flags);
+                }
+                /*
+                 * A number of cases require an RDS header to be sent
+                 * even if there is no data.
+                 * We permit 0-byte sends; rds-ping depends on this.
+                 * However, if there are exclusively attached silent ops,
+                 * we skip the hdr/data send, to enable silent operation.
+                 */
+                if (rm->data.op_nents == 0) {
+                        int ops_present;
+                        int all_ops_are_silent = 1;
+                        ops_present = (rm->atomic.op_active || rm->rdma.op_active);
+                        if (rm->atomic.op_active && !rm->atomic.op_silent)
+                                all_ops_are_silent = 0;
+                        if (rm->rdma.op_active && !rm->rdma.op_silent)
+                                all_ops_are_silent = 0;
+                        if (ops_present && all_ops_are_silent
+                            && !rm->m_rdma_cookie)
+                                rm->data.op_active = 0;
+                }
+                if (rm->data.op_active && !conn->c_xmit_data_sent) {
+                        rm->m_final_op = &rm->data;
                        ret = conn->c_trans->xmit(conn, rm,
                                                  conn->c_xmit_hdr_off,
                                                  conn->c_xmit_sg,
@@ -295,7 +314,7 @@ int rds_send_xmit(struct rds_connection *conn)
                                ret -= tmp;
                        }
-                        sg = &rm->m_sg[conn->c_xmit_sg];
+                        sg = &rm->data.op_sg[conn->c_xmit_sg];
                        while (ret) {
                                tmp = min_t(int, ret, sg->length -
                                                      conn->c_xmit_data_off);
@@ -306,49 +325,63 @@ int rds_send_xmit(struct rds_connection *conn)
                                        sg++;
                                        conn->c_xmit_sg++;
                                        BUG_ON(ret != 0 &&
-                                               conn->c_xmit_sg == rm->m_nents);
+                                               conn->c_xmit_sg == rm->data.op_nents);
                                }
                        }
+                        if (conn->c_xmit_hdr_off == sizeof(struct rds_header) &&
+                            (conn->c_xmit_sg == rm->data.op_nents))
+                                conn->c_xmit_data_sent = 1;
                }
-        }
-        /* Nuke any messages we decided not to retransmit. */
+                /*
-        if (!list_empty(&to_be_dropped))
+                 * A rm will only take multiple times through this loop
-                rds_send_remove_from_sock(&to_be_dropped, RDS_RDMA_DROPPED);
+                 * if there is a data op. Thus, if the data is sent (or there was
+                 * none), then we're done with the rm.
+                 */
+                if (!rm->data.op_active || conn->c_xmit_data_sent) {
+                        conn->c_xmit_rm = NULL;
+                        conn->c_xmit_sg = 0;
+                        conn->c_xmit_hdr_off = 0;
+                        conn->c_xmit_data_off = 0;
+                        conn->c_xmit_rdma_sent = 0;
+                        conn->c_xmit_atomic_sent = 0;
+                        conn->c_xmit_data_sent = 0;
+                        rds_message_put(rm);
+                }
+        }
        if (conn->c_trans->xmit_complete)
                conn->c_trans->xmit_complete(conn);
-        /*
+        release_in_xmit(conn);
-         * We might be racing with another sender who queued a message but
-         * backed off on noticing that we held the c_send_lock.  If we check
-         * for queued messages after dropping the sem then either we'll
-         * see the queued message or the queuer will get the sem.  If we
-         * notice the queued message then we trigger an immediate retry.
-         *
-         * We need to be careful only to do this when we stopped processing
-         * the send queue because it was empty.  It's the only way we
-         * stop processing the loop when the transport hasn't taken
-         * responsibility for forward progress.
-         */
-        mutex_unlock(&conn->c_send_lock);
-        if (conn->c_map_bytes || (send_quota == 0 && !was_empty)) {
+        /* Nuke any messages we decided not to retransmit. */
-                /* We exhausted the send quota, but there's work left to
+        if (!list_empty(&to_be_dropped)) {
-                 * do. Return and (re-)schedule the send worker.
+                /* irqs on here, so we can put(), unlike above */
-                 */
+                list_for_each_entry(rm, &to_be_dropped, m_conn_item)
-                ret = -EAGAIN;
+                        rds_message_put(rm);
+                rds_send_remove_from_sock(&to_be_dropped, RDS_RDMA_DROPPED);
        }
-        if (ret == 0 && was_empty) {
+        /*
-                /* A simple bit test would be way faster than taking the
+         * Other senders can queue a message after we last test the send queue
-                 * spin lock */
+         * but before we clear RDS_IN_XMIT.  In that case they'd back off and
-                spin_lock_irqsave(&conn->c_lock, flags);
+         * not try and send their newly queued message.  We need to check the
+         * send queue after having cleared RDS_IN_XMIT so that their message
+         * doesn't get stuck on the send queue.
+         *
+         * If the transport cannot continue (i.e ret != 0), then it must
+         * call us when more room is available, such as from the tx
+         * completion handler.
+         */
+        if (ret == 0) {
+                smp_mb();
                if (!list_empty(&conn->c_send_queue)) {
-                        rds_stats_inc(s_send_sem_queue_raced);
+                        rds_stats_inc(s_send_lock_queue_raced);
-                        ret = -EAGAIN;
+                        goto restart;
                }
-                spin_unlock_irqrestore(&conn->c_lock, flags);
        }
 out:
        return ret;
@@ -376,52 +409,60 @@ static inline int rds_send_is_acked(struct rds_message *rm, u64 ack,
 }
 /*
- * Returns true if there are no messages on the send and retransmit queues
+ * This is pretty similar to what happens below in the ACK
- * which have a sequence number greater than or equal to the given sequence
+ * handling code - except that we call here as soon as we get
- * number.
+ * the IB send completion on the RDMA op and the accompanying
+ * message.
 */
-int rds_send_acked_before(struct rds_connection *conn, u64 seq)
+void rds_rdma_send_complete(struct rds_message *rm, int status)
 {
-        struct rds_message *rm, *tmp;
+        struct rds_sock *rs = NULL;
-        int ret = 1;
+        struct rm_rdma_op *ro;
+        struct rds_notifier *notifier;
+        unsigned long flags;
-        spin_lock(&conn->c_lock);
+        spin_lock_irqsave(&rm->m_rs_lock, flags);
-        list_for_each_entry_safe(rm, tmp, &conn->c_retrans, m_conn_item) {
+        ro = &rm->rdma;
-                if (be64_to_cpu(rm->m_inc.i_hdr.h_sequence) < seq)
+        if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) &&
-                        ret = 0;
+            ro->op_active && ro->op_notify && ro->op_notifier) {
-                break;
+                notifier = ro->op_notifier;
-        }
+                rs = rm->m_rs;
+                sock_hold(rds_rs_to_sk(rs));
-        list_for_each_entry_safe(rm, tmp, &conn->c_send_queue, m_conn_item) {
+                notifier->n_status = status;
-                if (be64_to_cpu(rm->m_inc.i_hdr.h_sequence) < seq)
+                spin_lock(&rs->rs_lock);
-                        ret = 0;
+                list_add_tail(&notifier->n_list, &rs->rs_notify_queue);
-                break;
+                spin_unlock(&rs->rs_lock);
+                ro->op_notifier = NULL;
        }
-        spin_unlock(&conn->c_lock);
+        spin_unlock_irqrestore(&rm->m_rs_lock, flags);
-        return ret;
+        if (rs) {
+                rds_wake_sk_sleep(rs);
+                sock_put(rds_rs_to_sk(rs));
+        }
 }
+EXPORT_SYMBOL_GPL(rds_rdma_send_complete);
 /*
- * This is pretty similar to what happens below in the ACK
+ * Just like above, except looks at atomic op
- * handling code - except that we call here as soon as we get
- * the IB send completion on the RDMA op and the accompanying
- * message.
 */
-void rds_rdma_send_complete(struct rds_message *rm, int status)
+void rds_atomic_send_complete(struct rds_message *rm, int status)
 {
        struct rds_sock *rs = NULL;
-        struct rds_rdma_op *ro;
+        struct rm_atomic_op *ao;
        struct rds_notifier *notifier;
+        unsigned long flags;
-        spin_lock(&rm->m_rs_lock);
+        spin_lock_irqsave(&rm->m_rs_lock, flags);
-        ro = rm->m_rdma_op;
+        ao = &rm->atomic;
-        if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) &&
+        if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags)
-            ro && ro->r_notify && ro->r_notifier) {
+            && ao->op_active && ao->op_notify && ao->op_notifier) {
-                notifier = ro->r_notifier;
+                notifier = ao->op_notifier;
                rs = rm->m_rs;
                sock_hold(rds_rs_to_sk(rs));
@@ -430,17 +471,17 @@ void rds_rdma_send_complete(struct rds_message *rm, int status)
                list_add_tail(&notifier->n_list, &rs->rs_notify_queue);
                spin_unlock(&rs->rs_lock);
-                ro->r_notifier = NULL;
+                ao->op_notifier = NULL;
        }
-        spin_unlock(&rm->m_rs_lock);
+        spin_unlock_irqrestore(&rm->m_rs_lock, flags);
        if (rs) {
                rds_wake_sk_sleep(rs);
                sock_put(rds_rs_to_sk(rs));
        }
 }
-EXPORT_SYMBOL_GPL(rds_rdma_send_complete);
+EXPORT_SYMBOL_GPL(rds_atomic_send_complete);
 /*
 * This is the same as rds_rdma_send_complete except we
@@ -448,15 +489,23 @@ EXPORT_SYMBOL_GPL(rds_rdma_send_complete);
 * socket, socket lock) and can just move the notifier.
 */
 static inline void
-__rds_rdma_send_complete(struct rds_sock *rs, struct rds_message *rm, int status)
+__rds_send_complete(struct rds_sock *rs, struct rds_message *rm, int status)
 {
-        struct rds_rdma_op *ro;
+        struct rm_rdma_op *ro;
+        struct rm_atomic_op *ao;
+        ro = &rm->rdma;
+        if (ro->op_active && ro->op_notify && ro->op_notifier) {
+                ro->op_notifier->n_status = status;
+                list_add_tail(&ro->op_notifier->n_list, &rs->rs_notify_queue);
+                ro->op_notifier = NULL;
+        }
-        ro = rm->m_rdma_op;
+        ao = &rm->atomic;
-        if (ro && ro->r_notify && ro->r_notifier) {
+        if (ao->op_active && ao->op_notify && ao->op_notifier) {
-                ro->r_notifier->n_status = status;
+                ao->op_notifier->n_status = status;
-                list_add_tail(&ro->r_notifier->n_list, &rs->rs_notify_queue);
+                list_add_tail(&ao->op_notifier->n_list, &rs->rs_notify_queue);
-                ro->r_notifier = NULL;
+                ao->op_notifier = NULL;
        }
        /* No need to wake the app - caller does this */
@@ -468,7 +517,7 @@ __rds_rdma_send_complete(struct rds_sock *rs, struct rds_message *rm, int status
 * So speed is not an issue here.
 */
 struct rds_message *rds_send_get_message(struct rds_connection *conn,
-                                         struct rds_rdma_op *op)
+                                         struct rm_rdma_op *op)
 {
        struct rds_message *rm, *tmp, *found = NULL;
        unsigned long flags;
@@ -476,7 +525,7 @@ struct rds_message *rds_send_get_message(struct rds_connection *conn,
        spin_lock_irqsave(&conn->c_lock, flags);
        list_for_each_entry_safe(rm, tmp, &conn->c_retrans, m_conn_item) {
-                if (rm->m_rdma_op == op) {
+                if (&rm->rdma == op) {
                        atomic_inc(&rm->m_refcount);
                        found = rm;
                        goto out;
@@ -484,7 +533,7 @@ struct rds_message *rds_send_get_message(struct rds_connection *conn,
        }
        list_for_each_entry_safe(rm, tmp, &conn->c_send_queue, m_conn_item) {
-                if (rm->m_rdma_op == op) {
+                if (&rm->rdma == op) {
                        atomic_inc(&rm->m_refcount);
                        found = rm;
                        break;
@@ -544,19 +593,20 @@ void rds_send_remove_from_sock(struct list_head *messages, int status)
                spin_lock(&rs->rs_lock);
                if (test_and_clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags)) {
-                        struct rds_rdma_op *ro = rm->m_rdma_op;
+                        struct rm_rdma_op *ro = &rm->rdma;
                        struct rds_notifier *notifier;
                        list_del_init(&rm->m_sock_item);
                        rds_send_sndbuf_remove(rs, rm);
-                        if (ro && ro->r_notifier && (status || ro->r_notify)) {
+                        if (ro->op_active && ro->op_notifier &&
-                                notifier = ro->r_notifier;
+                               (ro->op_notify || (ro->op_recverr && status))) {
+                                notifier = ro->op_notifier;
                                list_add_tail(&notifier->n_list,
                                                &rs->rs_notify_queue);
                                if (!notifier->n_status)
                                        notifier->n_status = status;
-                                rm->m_rdma_op->r_notifier = NULL;
+                                rm->rdma.op_notifier = NULL;
                        }
                        was_on_sock = 1;
                        rm->m_rs = NULL;
@@ -619,9 +669,8 @@ void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in *dest)
 {
        struct rds_message *rm, *tmp;
        struct rds_connection *conn;
-        unsigned long flags, flags2;
+        unsigned long flags;
        LIST_HEAD(list);
-        int wake = 0;
        /* get all the messages we're dropping under the rs lock */
        spin_lock_irqsave(&rs->rs_lock, flags);
@@ -631,59 +680,54 @@ void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in *dest)
                             dest->sin_port != rm->m_inc.i_hdr.h_dport))
                        continue;
-                wake = 1;
                list_move(&rm->m_sock_item, &list);
                rds_send_sndbuf_remove(rs, rm);
                clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags);
        }
        /* order flag updates with the rs lock */
-        if (wake)
+        smp_mb__after_clear_bit();
-                smp_mb__after_clear_bit();
        spin_unlock_irqrestore(&rs->rs_lock, flags);
-        conn = NULL;
+        if (list_empty(&list))
+                return;
-        /* now remove the messages from the conn list as needed */
+        /* Remove the messages from the conn */
        list_for_each_entry(rm, &list, m_sock_item) {
-                /* We do this here rather than in the loop above, so that
-                 * we don't have to nest m_rs_lock under rs->rs_lock */
-                spin_lock_irqsave(&rm->m_rs_lock, flags2);
-                /* If this is a RDMA operation, notify the app. */
-                spin_lock(&rs->rs_lock);
-                __rds_rdma_send_complete(rs, rm, RDS_RDMA_CANCELED);
-                spin_unlock(&rs->rs_lock);
-                rm->m_rs = NULL;
-                spin_unlock_irqrestore(&rm->m_rs_lock, flags2);
+                conn = rm->m_inc.i_conn;
+                spin_lock_irqsave(&conn->c_lock, flags);
                /*
-                 * If we see this flag cleared then we're *sure* that someone
+                 * Maybe someone else beat us to removing rm from the conn.
-                 * else beat us to removing it from the conn.  If we race
+                 * If we race with their flag update we'll get the lock and
-                 * with their flag update we'll get the lock and then really
+                 * then really see that the flag has been cleared.
-                 * see that the flag has been cleared.
                 */
-                if (!test_bit(RDS_MSG_ON_CONN, &rm->m_flags))
+                if (!test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags)) {
+                        spin_unlock_irqrestore(&conn->c_lock, flags);
                        continue;
-                if (conn != rm->m_inc.i_conn) {
-                        if (conn)
-                                spin_unlock_irqrestore(&conn->c_lock, flags);
-                        conn = rm->m_inc.i_conn;
-                        spin_lock_irqsave(&conn->c_lock, flags);
                }
+                list_del_init(&rm->m_conn_item);
+                spin_unlock_irqrestore(&conn->c_lock, flags);
-                if (test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags)) {
+                /*
-                        list_del_init(&rm->m_conn_item);
+                 * Couldn't grab m_rs_lock in top loop (lock ordering),
-                        rds_message_put(rm);
+                 * but we can now.
-                }
+                 */
-        }
+                spin_lock_irqsave(&rm->m_rs_lock, flags);
-        if (conn)
+                spin_lock(&rs->rs_lock);
-                spin_unlock_irqrestore(&conn->c_lock, flags);
+                __rds_send_complete(rs, rm, RDS_RDMA_CANCELED);
+                spin_unlock(&rs->rs_lock);
-        if (wake)
+                rm->m_rs = NULL;
-                rds_wake_sk_sleep(rs);
+                spin_unlock_irqrestore(&rm->m_rs_lock, flags);
+                rds_message_put(rm);
+        }
+        rds_wake_sk_sleep(rs);
        while (!list_empty(&list)) {
                rm = list_entry(list.next, struct rds_message, m_sock_item);
@@ -763,6 +807,63 @@ out:
        return *queued;
 }
+/*
+ * rds_message is getting to be quite complicated, and we'd like to allocate
+ * it all in one go. This figures out how big it needs to be up front.
+ */
+static int rds_rm_size(struct msghdr *msg, int data_len)
+{
+        struct cmsghdr *cmsg;
+        int size = 0;
+        int cmsg_groups = 0;
+        int retval;
+        for (cmsg = CMSG_FIRSTHDR(msg); cmsg; cmsg = CMSG_NXTHDR(msg, cmsg)) {
+                if (!CMSG_OK(msg, cmsg))
+                        return -EINVAL;
+                if (cmsg->cmsg_level != SOL_RDS)
+                        continue;
+                switch (cmsg->cmsg_type) {
+                case RDS_CMSG_RDMA_ARGS:
+                        cmsg_groups |= 1;
+                        retval = rds_rdma_extra_size(CMSG_DATA(cmsg));
+                        if (retval < 0)
+                                return retval;
+                        size += retval;
+                        break;
+                case RDS_CMSG_RDMA_DEST:
+                case RDS_CMSG_RDMA_MAP:
+                        cmsg_groups |= 2;
+                        /* these are valid but do no add any size */
+                        break;
+                case RDS_CMSG_ATOMIC_CSWP:
+                case RDS_CMSG_ATOMIC_FADD:
+                case RDS_CMSG_MASKED_ATOMIC_CSWP:
+                case RDS_CMSG_MASKED_ATOMIC_FADD:
+                        cmsg_groups |= 1;
+                        size += sizeof(struct scatterlist);
+                        break;
+                default:
+                        return -EINVAL;
+                }
+        }
+        size += ceil(data_len, PAGE_SIZE) * sizeof(struct scatterlist);
+        /* Ensure (DEST, MAP) are never used with (ARGS, ATOMIC) */
+        if (cmsg_groups == 3)
+                return -EINVAL;
+        return size;
+}
 static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
                         struct msghdr *msg, int *allocated_mr)
 {
@@ -777,7 +878,7 @@ static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
                        continue;
                /* As a side effect, RDMA_DEST and RDMA_MAP will set
-                 * rm->m_rdma_cookie and rm->m_rdma_mr.
+                 * rm->rdma.m_rdma_cookie and rm->rdma.m_rdma_mr.
                 */
                switch (cmsg->cmsg_type) {
                case RDS_CMSG_RDMA_ARGS:
@@ -793,6 +894,12 @@ static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
                        if (!ret)
                                *allocated_mr = 1;
                        break;
+                case RDS_CMSG_ATOMIC_CSWP:
+                case RDS_CMSG_ATOMIC_FADD:
+                case RDS_CMSG_MASKED_ATOMIC_CSWP:
+                case RDS_CMSG_MASKED_ATOMIC_FADD:
+                        ret = rds_cmsg_atomic(rs, rm, cmsg);
+                        break;
                default:
                        return -EINVAL;
@@ -850,13 +957,26 @@ int rds_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
                goto out;
        }
-        rm = rds_message_copy_from_user(msg->msg_iov, payload_len);
+        /* size of rm including all sgs */
-        if (IS_ERR(rm)) {
+        ret = rds_rm_size(msg, payload_len);
-                ret = PTR_ERR(rm);
+        if (ret < 0)
-                rm = NULL;
+                goto out;
+        rm = rds_message_alloc(ret, GFP_KERNEL);
+        if (!rm) {
+                ret = -ENOMEM;
                goto out;
        }
+        /* Attach data to the rm */
+        if (payload_len) {
+                rm->data.op_sg = rds_message_alloc_sgs(rm, ceil(payload_len, PAGE_SIZE));
+                ret = rds_message_copy_from_user(rm, msg->msg_iov, payload_len);
+                if (ret)
+                        goto out;
+        }
+        rm->data.op_active = 1;
        rm->m_daddr = daddr;
        /* rds_conn_create has a spinlock that runs with IRQ off.
@@ -879,22 +999,23 @@ int rds_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
        if (ret)
                goto out;
-        if ((rm->m_rdma_cookie || rm->m_rdma_op) &&
+        if (rm->rdma.op_active && !conn->c_trans->xmit_rdma) {
-            conn->c_trans->xmit_rdma == NULL) {
                if (printk_ratelimit())
                        printk(KERN_NOTICE "rdma_op %p conn xmit_rdma %p\n",
-                                rm->m_rdma_op, conn->c_trans->xmit_rdma);
+                               &rm->rdma, conn->c_trans->xmit_rdma);
                ret = -EOPNOTSUPP;
                goto out;
        }
-        /* If the connection is down, trigger a connect. We may
+        if (rm->atomic.op_active && !conn->c_trans->xmit_atomic) {
-         * have scheduled a delayed reconnect however - in this case
+                if (printk_ratelimit())
-         * we should not interfere.
+                        printk(KERN_NOTICE "atomic_op %p conn xmit_atomic %p\n",
-         */
+                               &rm->atomic, conn->c_trans->xmit_atomic);
-        if (rds_conn_state(conn) == RDS_CONN_DOWN &&
+                ret = -EOPNOTSUPP;
-            !test_and_set_bit(RDS_RECONNECT_PENDING, &conn->c_flags))
+                goto out;
-                queue_delayed_work(rds_wq, &conn->c_conn_w, 0);
+        }
+        rds_conn_connect_if_down(conn);
        ret = rds_cong_wait(conn->c_fcong, dport, nonblock, rs);
        if (ret) {
@@ -938,7 +1059,7 @@ int rds_sendmsg(struct kiocb *iocb, struct socket *sock, struct msghdr *msg,
        rds_stats_inc(s_send_queued);
        if (!test_bit(RDS_LL_SEND_FULL, &conn->c_flags))
-                rds_send_worker(&conn->c_send_w.work);
+                rds_send_xmit(conn);
        rds_message_put(rm);
        return payload_len;
@@ -966,20 +1087,15 @@ rds_send_pong(struct rds_connection *conn, __be16 dport)
        int ret = 0;
        rm = rds_message_alloc(0, GFP_ATOMIC);
-        if (rm == NULL) {
+        if (!rm) {
                ret = -ENOMEM;
                goto out;
        }
        rm->m_daddr = conn->c_faddr;
+        rm->data.op_active = 1;
-        /* If the connection is down, trigger a connect. We may
+        rds_conn_connect_if_down(conn);
-         * have scheduled a delayed reconnect however - in this case
-         * we should not interfere.
-         */
-        if (rds_conn_state(conn) == RDS_CONN_DOWN &&
-            !test_and_set_bit(RDS_RECONNECT_PENDING, &conn->c_flags))
-                queue_delayed_work(rds_wq, &conn->c_conn_w, 0);
        ret = rds_cong_wait(conn->c_fcong, dport, 1, NULL);
        if (ret)
@@ -999,7 +1115,9 @@ rds_send_pong(struct rds_connection *conn, __be16 dport)
        rds_stats_inc(s_send_queued);
        rds_stats_inc(s_send_pong);
-        queue_delayed_work(rds_wq, &conn->c_send_w, 0);
+        if (!test_bit(RDS_LL_SEND_FULL, &conn->c_flags))
+                rds_send_xmit(conn);
        rds_message_put(rm);
        return 0;
diff --git a/net/rds/stats.c b/net/rds/stats.c
index 7598eb07cfb1..10c759ccac0c 100644
--- a/net/rds/stats.c
+++ b/net/rds/stats.c
@@ -57,8 +57,8 @@ static const char *const rds_stat_names[] = {
        "recv_ping",
        "send_queue_empty",
        "send_queue_full",
-        "send_sem_contention",
+        "send_lock_contention",
-        "send_sem_queue_raced",
+        "send_lock_queue_raced",
        "send_immediate_retry",
        "send_delayed_retry",
        "send_drop_acked",
@@ -143,7 +143,7 @@ void rds_stats_exit(void)
        rds_info_deregister_func(RDS_INFO_COUNTERS, rds_stats_info);
 }
-int __init rds_stats_init(void)
+int rds_stats_init(void)
 {
        rds_info_register_func(RDS_INFO_COUNTERS, rds_stats_info);
        return 0;
diff --git a/net/rds/sysctl.c b/net/rds/sysctl.c
index 7829a20325d3..25ad0c77a26c 100644
--- a/net/rds/sysctl.c
+++ b/net/rds/sysctl.c
@@ -105,13 +105,13 @@ void rds_sysctl_exit(void)
                unregister_sysctl_table(rds_sysctl_reg_table);
 }
-int __init rds_sysctl_init(void)
+int rds_sysctl_init(void)
 {
        rds_sysctl_reconnect_min = msecs_to_jiffies(1);
        rds_sysctl_reconnect_min_jiffies = rds_sysctl_reconnect_min;
        rds_sysctl_reg_table = register_sysctl_paths(rds_sysctl_path, rds_sysctl_rds_table);
-        if (rds_sysctl_reg_table == NULL)
+        if (!rds_sysctl_reg_table)
                return -ENOMEM;
        return 0;
 }
diff --git a/net/rds/tcp.c b/net/rds/tcp.c
index babf4577ff7d..eeb08e6ab96b 100644
--- a/net/rds/tcp.c
+++ b/net/rds/tcp.c
@@ -200,7 +200,7 @@ static int rds_tcp_conn_alloc(struct rds_connection *conn, gfp_t gfp)
        struct rds_tcp_connection *tc;
        tc = kmem_cache_alloc(rds_tcp_conn_slab, gfp);
-        if (tc == NULL)
+        if (!tc)
                return -ENOMEM;
        tc->t_sock = NULL;
@@ -258,7 +258,6 @@ struct rds_transport rds_tcp_transport = {
        .laddr_check            = rds_tcp_laddr_check,
        .xmit_prepare           = rds_tcp_xmit_prepare,
        .xmit_complete          = rds_tcp_xmit_complete,
-        .xmit_cong_map          = rds_tcp_xmit_cong_map,
        .xmit                   = rds_tcp_xmit,
        .recv                   = rds_tcp_recv,
        .conn_alloc             = rds_tcp_conn_alloc,
@@ -266,7 +265,6 @@ struct rds_transport rds_tcp_transport = {
        .conn_connect           = rds_tcp_conn_connect,
        .conn_shutdown          = rds_tcp_conn_shutdown,
        .inc_copy_to_user       = rds_tcp_inc_copy_to_user,
-        .inc_purge              = rds_tcp_inc_purge,
        .inc_free               = rds_tcp_inc_free,
        .stats_info_copy        = rds_tcp_stats_info_copy,
        .exit                   = rds_tcp_exit,
@@ -276,14 +274,14 @@ struct rds_transport rds_tcp_transport = {
        .t_prefer_loopback      = 1,
 };
-int __init rds_tcp_init(void)
+int rds_tcp_init(void)
 {
        int ret;
        rds_tcp_conn_slab = kmem_cache_create("rds_tcp_connection",
                                              sizeof(struct rds_tcp_connection),
                                              0, 0, NULL);
-        if (rds_tcp_conn_slab == NULL) {
+        if (!rds_tcp_conn_slab) {
                ret = -ENOMEM;
                goto out;
        }
diff --git a/net/rds/tcp.h b/net/rds/tcp.h
index 844fa6b9cf5a..f5e6f7bebb50 100644
--- a/net/rds/tcp.h
+++ b/net/rds/tcp.h
@@ -43,7 +43,7 @@ struct rds_tcp_statistics {
 };
 /* tcp.c */
-int __init rds_tcp_init(void);
+int rds_tcp_init(void);
 void rds_tcp_exit(void);
 void rds_tcp_tune(struct socket *sock);
 void rds_tcp_nonagle(struct socket *sock);
@@ -61,16 +61,15 @@ void rds_tcp_conn_shutdown(struct rds_connection *conn);
 void rds_tcp_state_change(struct sock *sk);
 /* tcp_listen.c */
-int __init rds_tcp_listen_init(void);
+int rds_tcp_listen_init(void);
 void rds_tcp_listen_stop(void);
 void rds_tcp_listen_data_ready(struct sock *sk, int bytes);
 /* tcp_recv.c */
-int __init rds_tcp_recv_init(void);
+int rds_tcp_recv_init(void);
 void rds_tcp_recv_exit(void);
 void rds_tcp_data_ready(struct sock *sk, int bytes);
 int rds_tcp_recv(struct rds_connection *conn);
-void rds_tcp_inc_purge(struct rds_incoming *inc);
 void rds_tcp_inc_free(struct rds_incoming *inc);
 int rds_tcp_inc_copy_to_user(struct rds_incoming *inc, struct iovec *iov,
                             size_t size);
@@ -81,8 +80,6 @@ void rds_tcp_xmit_complete(struct rds_connection *conn);
 int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
                 unsigned int hdr_off, unsigned int sg, unsigned int off);
 void rds_tcp_write_space(struct sock *sk);
-int rds_tcp_xmit_cong_map(struct rds_connection *conn,
-                          struct rds_cong_map *map, unsigned long offset);
 /* tcp_stats.c */
 DECLARE_PER_CPU(struct rds_tcp_statistics, rds_tcp_stats);
diff --git a/net/rds/tcp_connect.c b/net/rds/tcp_connect.c
index c519939e8da9..af95c8e058fc 100644
--- a/net/rds/tcp_connect.c
+++ b/net/rds/tcp_connect.c
@@ -45,7 +45,7 @@ void rds_tcp_state_change(struct sock *sk)
        read_lock_bh(&sk->sk_callback_lock);
        conn = sk->sk_user_data;
-        if (conn == NULL) {
+        if (!conn) {
                state_change = sk->sk_state_change;
                goto out;
        }
diff --git a/net/rds/tcp_listen.c b/net/rds/tcp_listen.c
index 27844f231d10..8b5cc4aa8868 100644
--- a/net/rds/tcp_listen.c
+++ b/net/rds/tcp_listen.c
@@ -116,7 +116,7 @@ void rds_tcp_listen_data_ready(struct sock *sk, int bytes)
        read_lock_bh(&sk->sk_callback_lock);
        ready = sk->sk_user_data;
-        if (ready == NULL) { /* check for teardown race */
+        if (!ready) { /* check for teardown race */
                ready = sk->sk_data_ready;
                goto out;
        }
@@ -135,7 +135,7 @@ out:
        ready(sk, bytes);
 }
-int __init rds_tcp_listen_init(void)
+int rds_tcp_listen_init(void)
 {
        struct sockaddr_in sin;
        struct socket *sock = NULL;
@@ -178,7 +178,7 @@ void rds_tcp_listen_stop(void)
        struct socket *sock = rds_tcp_listen_sock;
        struct sock *sk;
-        if (sock == NULL)
+        if (!sock)
                return;
        sk = sock->sk;
diff --git a/net/rds/tcp_recv.c b/net/rds/tcp_recv.c
index e43797404102..67263fbee623 100644
--- a/net/rds/tcp_recv.c
+++ b/net/rds/tcp_recv.c
@@ -39,7 +39,7 @@
 static struct kmem_cache *rds_tcp_incoming_slab;
-void rds_tcp_inc_purge(struct rds_incoming *inc)
+static void rds_tcp_inc_purge(struct rds_incoming *inc)
 {
        struct rds_tcp_incoming *tinc;
        tinc = container_of(inc, struct rds_tcp_incoming, ti_inc);
@@ -190,10 +190,10 @@ static int rds_tcp_data_recv(read_descriptor_t *desc, struct sk_buff *skb,
         * processing.
         */
        while (left) {
-                if (tinc == NULL) {
+                if (!tinc) {
                        tinc = kmem_cache_alloc(rds_tcp_incoming_slab,
                                                arg->gfp);
-                        if (tinc == NULL) {
+                        if (!tinc) {
                                desc->error = -ENOMEM;
                                goto out;
                        }
@@ -229,7 +229,7 @@ static int rds_tcp_data_recv(read_descriptor_t *desc, struct sk_buff *skb,
                if (left && tc->t_tinc_data_rem) {
                        clone = skb_clone(skb, arg->gfp);
-                        if (clone == NULL) {
+                        if (!clone) {
                                desc->error = -ENOMEM;
                                goto out;
                        }
@@ -326,7 +326,7 @@ void rds_tcp_data_ready(struct sock *sk, int bytes)
        read_lock_bh(&sk->sk_callback_lock);
        conn = sk->sk_user_data;
-        if (conn == NULL) { /* check for teardown race */
+        if (!conn) { /* check for teardown race */
                ready = sk->sk_data_ready;
                goto out;
        }
@@ -342,12 +342,12 @@ out:
        ready(sk, bytes);
 }
-int __init rds_tcp_recv_init(void)
+int rds_tcp_recv_init(void)
 {
        rds_tcp_incoming_slab = kmem_cache_create("rds_tcp_incoming",
                                        sizeof(struct rds_tcp_incoming),
                                        0, 0, NULL);
-        if (rds_tcp_incoming_slab == NULL)
+        if (!rds_tcp_incoming_slab)
                return -ENOMEM;
        return 0;
 }
diff --git a/net/rds/tcp_send.c b/net/rds/tcp_send.c
index 2f012a07d94d..aa16841afbdf 100644
--- a/net/rds/tcp_send.c
+++ b/net/rds/tcp_send.c
@@ -77,56 +77,6 @@ int rds_tcp_sendmsg(struct socket *sock, void *data, unsigned int len)
 }
 /* the core send_sem serializes this with other xmit and shutdown */
-int rds_tcp_xmit_cong_map(struct rds_connection *conn,
-                          struct rds_cong_map *map, unsigned long offset)
-{
-        static struct rds_header rds_tcp_map_header = {
-                .h_flags = RDS_FLAG_CONG_BITMAP,
-        };
-        struct rds_tcp_connection *tc = conn->c_transport_data;
-        unsigned long i;
-        int ret;
-        int copied = 0;
-        /* Some problem claims cpu_to_be32(constant) isn't a constant. */
-        rds_tcp_map_header.h_len = cpu_to_be32(RDS_CONG_MAP_BYTES);
-        if (offset < sizeof(struct rds_header)) {
-                ret = rds_tcp_sendmsg(tc->t_sock,
-                                      (void *)&rds_tcp_map_header + offset,
-                                      sizeof(struct rds_header) - offset);
-                if (ret <= 0)
-                        return ret;
-                offset += ret;
-                copied = ret;
-                if (offset < sizeof(struct rds_header))
-                        return ret;
-        }
-        offset -= sizeof(struct rds_header);
-        i = offset / PAGE_SIZE;
-        offset = offset % PAGE_SIZE;
-        BUG_ON(i >= RDS_CONG_MAP_PAGES);
-        do {
-                ret = tc->t_sock->ops->sendpage(tc->t_sock,
-                                        virt_to_page(map->m_page_addrs[i]),
-                                        offset, PAGE_SIZE - offset,
-                                        MSG_DONTWAIT);
-                if (ret <= 0)
-                        break;
-                copied += ret;
-                offset += ret;
-                if (offset == PAGE_SIZE) {
-                        offset = 0;
-                        i++;
-                }
-        } while (i < RDS_CONG_MAP_PAGES);
-        return copied ? copied : ret;
-}
-/* the core send_sem serializes this with other xmit and shutdown */
 int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
                 unsigned int hdr_off, unsigned int sg, unsigned int off)
 {
@@ -166,21 +116,21 @@ int rds_tcp_xmit(struct rds_connection *conn, struct rds_message *rm,
                        goto out;
        }
-        while (sg < rm->m_nents) {
+        while (sg < rm->data.op_nents) {
                ret = tc->t_sock->ops->sendpage(tc->t_sock,
-                                                sg_page(&rm->m_sg[sg]),
+                                                sg_page(&rm->data.op_sg[sg]),
-                                                rm->m_sg[sg].offset + off,
+                                                rm->data.op_sg[sg].offset + off,
-                                                rm->m_sg[sg].length - off,
+                                                rm->data.op_sg[sg].length - off,
                                                MSG_DONTWAIT|MSG_NOSIGNAL);
-                rdsdebug("tcp sendpage %p:%u:%u ret %d\n", (void *)sg_page(&rm->m_sg[sg]),
+                rdsdebug("tcp sendpage %p:%u:%u ret %d\n", (void *)sg_page(&rm->data.op_sg[sg]),
-                         rm->m_sg[sg].offset + off, rm->m_sg[sg].length - off,
+                         rm->data.op_sg[sg].offset + off, rm->data.op_sg[sg].length - off,
                         ret);
                if (ret <= 0)
                        break;
                off += ret;
                done += ret;
-                if (off == rm->m_sg[sg].length) {
+                if (off == rm->data.op_sg[sg].length) {
                        off = 0;
                        sg++;
                }
@@ -226,7 +176,7 @@ void rds_tcp_write_space(struct sock *sk)
        read_lock_bh(&sk->sk_callback_lock);
        conn = sk->sk_user_data;
-        if (conn == NULL) {
+        if (!conn) {
                write_space = sk->sk_write_space;
                goto out;
        }
diff --git a/net/rds/threads.c b/net/rds/threads.c
index 786c20eaaf5e..0fd90f8c5f59 100644
--- a/net/rds/threads.c
+++ b/net/rds/threads.c
@@ -61,7 +61,7 @@
 *
 * Transition to state DISCONNECTING/DOWN:
 *  -   Inside the shutdown worker; synchronizes with xmit path
- *      through c_send_lock, and with connection management callbacks
+ *      through RDS_IN_XMIT, and with connection management callbacks
 *      via c_cm_lock.
 *
 *      For receive callbacks, we rely on the underlying transport
@@ -110,7 +110,7 @@ EXPORT_SYMBOL_GPL(rds_connect_complete);
 * We should *always* start with a random backoff; otherwise a broken connection
 * will always take several iterations to be re-established.
 */
-static void rds_queue_reconnect(struct rds_connection *conn)
+void rds_queue_reconnect(struct rds_connection *conn)
 {
        unsigned long rand;
@@ -156,58 +156,6 @@ void rds_connect_worker(struct work_struct *work)
        }
 }
-void rds_shutdown_worker(struct work_struct *work)
-{
-        struct rds_connection *conn = container_of(work, struct rds_connection, c_down_w);
-        /* shut it down unless it's down already */
-        if (!rds_conn_transition(conn, RDS_CONN_DOWN, RDS_CONN_DOWN)) {
-                /*
-                 * Quiesce the connection mgmt handlers before we start tearing
-                 * things down. We don't hold the mutex for the entire
-                 * duration of the shutdown operation, else we may be
-                 * deadlocking with the CM handler. Instead, the CM event
-                 * handler is supposed to check for state DISCONNECTING
-                 */
-                mutex_lock(&conn->c_cm_lock);
-                if (!rds_conn_transition(conn, RDS_CONN_UP, RDS_CONN_DISCONNECTING) &&
-                    !rds_conn_transition(conn, RDS_CONN_ERROR, RDS_CONN_DISCONNECTING)) {
-                        rds_conn_error(conn, "shutdown called in state %d\n",
-                                        atomic_read(&conn->c_state));
-                        mutex_unlock(&conn->c_cm_lock);
-                        return;
-                }
-                mutex_unlock(&conn->c_cm_lock);
-                mutex_lock(&conn->c_send_lock);
-                conn->c_trans->conn_shutdown(conn);
-                rds_conn_reset(conn);
-                mutex_unlock(&conn->c_send_lock);
-                if (!rds_conn_transition(conn, RDS_CONN_DISCONNECTING, RDS_CONN_DOWN)) {
-                        /* This can happen - eg when we're in the middle of tearing
-                         * down the connection, and someone unloads the rds module.
-                         * Quite reproduceable with loopback connections.
-                         * Mostly harmless.
-                         */
-                        rds_conn_error(conn,
-                                "%s: failed to transition to state DOWN, "
-                                "current state is %d\n",
-                                __func__,
-                                atomic_read(&conn->c_state));
-                        return;
-                }
-        }
-        /* Then reconnect if it's still live.
-         * The passive side of an IB loopback connection is never added
-         * to the conn hash, so we never trigger a reconnect on this
-         * conn - the reconnect is always triggered by the active peer. */
-        cancel_delayed_work(&conn->c_conn_w);
-        if (!hlist_unhashed(&conn->c_hash_node))
-                rds_queue_reconnect(conn);
-}
 void rds_send_worker(struct work_struct *work)
 {
        struct rds_connection *conn = container_of(work, struct rds_connection, c_send_w.work);
@@ -252,15 +200,22 @@ void rds_recv_worker(struct work_struct *work)
        }
 }
+void rds_shutdown_worker(struct work_struct *work)
+{
+        struct rds_connection *conn = container_of(work, struct rds_connection, c_down_w);
+        rds_conn_shutdown(conn);
+}
 void rds_threads_exit(void)
 {
        destroy_workqueue(rds_wq);
 }
-int __init rds_threads_init(void)
+int rds_threads_init(void)
 {
-        rds_wq = create_workqueue("krdsd");
+        rds_wq = create_singlethread_workqueue("krdsd");
-        if (rds_wq == NULL)
+        if (!rds_wq)
                return -ENOMEM;
        return 0;
diff --git a/net/rds/transport.c b/net/rds/transport.c
index 7e1067901353..7f2ac4fec367 100644
--- a/net/rds/transport.c
+++ b/net/rds/transport.c
@@ -71,19 +71,28 @@ void rds_trans_unregister(struct rds_transport *trans)
 }
 EXPORT_SYMBOL_GPL(rds_trans_unregister);
+void rds_trans_put(struct rds_transport *trans)
+{
+        if (trans && trans->t_owner)
+                module_put(trans->t_owner);
+}
 struct rds_transport *rds_trans_get_preferred(__be32 addr)
 {
        struct rds_transport *ret = NULL;
-        int i;
+        struct rds_transport *trans;
+        unsigned int i;
        if (IN_LOOPBACK(ntohl(addr)))
                return &rds_loop_transport;
        down_read(&rds_trans_sem);
-        for (i = 0; i < RDS_TRANS_COUNT; i++)
+        for (i = 0; i < RDS_TRANS_COUNT; i++) {
-        {
+                trans = transports[i];
-                if (transports[i] && (transports[i]->laddr_check(addr) == 0)) {
-                        ret = transports[i];
+                if (trans && (trans->laddr_check(addr) == 0) &&
+                    (!trans->t_owner || try_module_get(trans->t_owner))) {
+                        ret = trans;
                        break;
                }
        }
diff --git a/net/rds/xlist.h b/net/rds/xlist.h
new file mode 100644
index 000000000000..e6b5190daddd
--- /dev/null
+++ b/net/rds/xlist.h
@@ -0,0 +1,80 @@
+#ifndef _LINUX_XLIST_H
+#define _LINUX_XLIST_H
+#include <linux/stddef.h>
+#include <linux/poison.h>
+#include <linux/prefetch.h>
+#include <asm/system.h>
+struct xlist_head {
+        struct xlist_head *next;
+};
+static inline void INIT_XLIST_HEAD(struct xlist_head *list)
+{
+        list->next = NULL;
+}
+static inline int xlist_empty(struct xlist_head *head)
+{
+        return head->next == NULL;
+}
+static inline void xlist_add(struct xlist_head *new, struct xlist_head *tail,
+                             struct xlist_head *head)
+{
+        struct xlist_head *cur;
+        struct xlist_head *check;
+        while (1) {
+                cur = head->next;
+                tail->next = cur;
+                check = cmpxchg(&head->next, cur, new);
+                if (check == cur)
+                        break;
+        }
+}
+static inline struct xlist_head *xlist_del_head(struct xlist_head *head)
+{
+        struct xlist_head *cur;
+        struct xlist_head *check;
+        struct xlist_head *next;
+        while (1) {
+                cur = head->next;
+                if (!cur)
+                        goto out;
+                next = cur->next;
+                check = cmpxchg(&head->next, cur, next);
+                if (check == cur)
+                        goto out;
+        }
+out:
+        return cur;
+}
+static inline struct xlist_head *xlist_del_head_fast(struct xlist_head *head)
+{
+        struct xlist_head *cur;
+        cur = head->next;
+        if (!cur)
+                return NULL;
+        head->next = cur->next;
+        return cur;
+}
+static inline void xlist_splice(struct xlist_head *list,
+                                struct xlist_head *head)
+{
+        struct xlist_head *cur;
+        WARN_ON(head->next);
+        cur = xchg(&list->next, NULL);
+        head->next = cur;
+}
+#endif
diff --git a/net/rfkill/input.c b/net/rfkill/input.c
index 3713d7ecab96..1bca6d49ec96 100644
--- a/net/rfkill/input.c
+++ b/net/rfkill/input.c
@@ -142,7 +142,7 @@ static unsigned long rfkill_last_scheduled;
 static unsigned long rfkill_ratelimit(const unsigned long last)
 {
        const unsigned long delay = msecs_to_jiffies(RFKILL_OPS_DELAY);
-        return (time_after(jiffies, last + delay)) ? 0 : delay;
+        return time_after(jiffies, last + delay) ? 0 : delay;
 }
 static void rfkill_schedule_ratelimited(void)
diff --git a/net/rose/rose_link.c b/net/rose/rose_link.c
index a750a28e0221..fa5f5641a2c2 100644
--- a/net/rose/rose_link.c
+++ b/net/rose/rose_link.c
@@ -114,7 +114,7 @@ static int rose_send_frame(struct sk_buff *skb, struct rose_neigh *neigh)
        if (ax25s)
                ax25_cb_put(ax25s);
-        return (neigh->ax25 != NULL);
+        return neigh->ax25 != NULL;
 }
 /*
@@ -137,7 +137,7 @@ static int rose_link_up(struct rose_neigh *neigh)
        if (ax25s)
                ax25_cb_put(ax25s);
-        return (neigh->ax25 != NULL);
+        return neigh->ax25 != NULL;
 }
 /*
diff --git a/net/sched/Kconfig b/net/sched/Kconfig
index 2f691fb180d1..a36270a994d7 100644
--- a/net/sched/Kconfig
+++ b/net/sched/Kconfig
@@ -518,6 +518,16 @@ config NET_ACT_SKBEDIT
          To compile this code as a module, choose M here: the
          module will be called act_skbedit.
+config NET_ACT_CSUM
+        tristate "Checksum Updating"
+        depends on NET_CLS_ACT && INET
+        ---help---
+          Say Y here to update some common checksum after some direct
+          packet alterations.
+          To compile this code as a module, choose M here: the
+          module will be called act_csum.
 config NET_CLS_IND
        bool "Incoming device classification"
        depends on NET_CLS_U32 || NET_CLS_FW
diff --git a/net/sched/Makefile b/net/sched/Makefile
index f14e71bfa58f..960f5dba6304 100644
--- a/net/sched/Makefile
+++ b/net/sched/Makefile
@@ -15,6 +15,7 @@ obj-$(CONFIG_NET_ACT_NAT)	+= act_nat.o
 obj-$(CONFIG_NET_ACT_PEDIT)     += act_pedit.o
 obj-$(CONFIG_NET_ACT_SIMP)      += act_simple.o
 obj-$(CONFIG_NET_ACT_SKBEDIT)   += act_skbedit.o
+obj-$(CONFIG_NET_ACT_CSUM)      += act_csum.o
 obj-$(CONFIG_NET_SCH_FIFO)      += sch_fifo.o
 obj-$(CONFIG_NET_SCH_CBQ)       += sch_cbq.o
 obj-$(CONFIG_NET_SCH_HTB)       += sch_htb.o
diff --git a/net/sched/act_csum.c b/net/sched/act_csum.c
new file mode 100644
index 000000000000..67dc7ce9b63a
--- /dev/null
+++ b/net/sched/act_csum.c
@@ -0,0 +1,595 @@
+/*
+ * Checksum updating actions
+ *
+ * Copyright (c) 2010 Gregoire Baron <baronchon@n7mm.org>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ */
+#include <linux/types.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/spinlock.h>
+#include <linux/netlink.h>
+#include <net/netlink.h>
+#include <linux/rtnetlink.h>
+#include <linux/skbuff.h>
+#include <net/ip.h>
+#include <net/ipv6.h>
+#include <net/icmp.h>
+#include <linux/icmpv6.h>
+#include <linux/igmp.h>
+#include <net/tcp.h>
+#include <net/udp.h>
+#include <net/ip6_checksum.h>
+#include <net/act_api.h>
+#include <linux/tc_act/tc_csum.h>
+#include <net/tc_act/tc_csum.h>
+#define CSUM_TAB_MASK 15
+static struct tcf_common *tcf_csum_ht[CSUM_TAB_MASK + 1];
+static u32 csum_idx_gen;
+static DEFINE_RWLOCK(csum_lock);
+static struct tcf_hashinfo csum_hash_info = {
+        .htab   = tcf_csum_ht,
+        .hmask  = CSUM_TAB_MASK,
+        .lock   = &csum_lock,
+};
+static const struct nla_policy csum_policy[TCA_CSUM_MAX + 1] = {
+        [TCA_CSUM_PARMS] = { .len = sizeof(struct tc_csum), },
+};
+static int tcf_csum_init(struct nlattr *nla, struct nlattr *est,
+                         struct tc_action *a, int ovr, int bind)
+{
+        struct nlattr *tb[TCA_CSUM_MAX + 1];
+        struct tc_csum *parm;
+        struct tcf_common *pc;
+        struct tcf_csum *p;
+        int ret = 0, err;
+        if (nla == NULL)
+                return -EINVAL;
+        err = nla_parse_nested(tb, TCA_CSUM_MAX, nla,csum_policy);
+        if (err < 0)
+                return err;
+        if (tb[TCA_CSUM_PARMS] == NULL)
+                return -EINVAL;
+        parm = nla_data(tb[TCA_CSUM_PARMS]);
+        pc = tcf_hash_check(parm->index, a, bind, &csum_hash_info);
+        if (!pc) {
+                pc = tcf_hash_create(parm->index, est, a, sizeof(*p), bind,
+                                     &csum_idx_gen, &csum_hash_info);
+                if (IS_ERR(pc))
+                        return PTR_ERR(pc);
+                p = to_tcf_csum(pc);
+                ret = ACT_P_CREATED;
+        } else {
+                p = to_tcf_csum(pc);
+                if (!ovr) {
+                        tcf_hash_release(pc, bind, &csum_hash_info);
+                        return -EEXIST;
+                }
+        }
+        spin_lock_bh(&p->tcf_lock);
+        p->tcf_action = parm->action;
+        p->update_flags = parm->update_flags;
+        spin_unlock_bh(&p->tcf_lock);
+        if (ret == ACT_P_CREATED)
+                tcf_hash_insert(pc, &csum_hash_info);
+        return ret;
+}
+static int tcf_csum_cleanup(struct tc_action *a, int bind)
+{
+        struct tcf_csum *p = a->priv;
+        return tcf_hash_release(&p->common, bind, &csum_hash_info);
+}
+/**
+ * tcf_csum_skb_nextlayer - Get next layer pointer
+ * @skb: sk_buff to use
+ * @ihl: previous summed headers length
+ * @ipl: complete packet length
+ * @jhl: next header length
+ *
+ * Check the expected next layer availability in the specified sk_buff.
+ * Return the next layer pointer if pass, NULL otherwise.
+ */
+static void *tcf_csum_skb_nextlayer(struct sk_buff *skb,
+                                    unsigned int ihl, unsigned int ipl,
+                                    unsigned int jhl)
+{
+        int ntkoff = skb_network_offset(skb);
+        int hl = ihl + jhl;
+        if (!pskb_may_pull(skb, ipl + ntkoff) || (ipl < hl) ||
+            (skb_cloned(skb) &&
+             !skb_clone_writable(skb, hl + ntkoff) &&
+             pskb_expand_head(skb, 0, 0, GFP_ATOMIC)))
+                return NULL;
+        else
+                return (void *)(skb_network_header(skb) + ihl);
+}
+static int tcf_csum_ipv4_icmp(struct sk_buff *skb,
+                              unsigned int ihl, unsigned int ipl)
+{
+        struct icmphdr *icmph;
+        icmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmph));
+        if (icmph == NULL)
+                return 0;
+        icmph->checksum = 0;
+        skb->csum = csum_partial(icmph, ipl - ihl, 0);
+        icmph->checksum = csum_fold(skb->csum);
+        skb->ip_summed = CHECKSUM_NONE;
+        return 1;
+}
+static int tcf_csum_ipv4_igmp(struct sk_buff *skb,
+                              unsigned int ihl, unsigned int ipl)
+{
+        struct igmphdr *igmph;
+        igmph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*igmph));
+        if (igmph == NULL)
+                return 0;
+        igmph->csum = 0;
+        skb->csum = csum_partial(igmph, ipl - ihl, 0);
+        igmph->csum = csum_fold(skb->csum);
+        skb->ip_summed = CHECKSUM_NONE;
+        return 1;
+}
+static int tcf_csum_ipv6_icmp(struct sk_buff *skb, struct ipv6hdr *ip6h,
+                              unsigned int ihl, unsigned int ipl)
+{
+        struct icmp6hdr *icmp6h;
+        icmp6h = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*icmp6h));
+        if (icmp6h == NULL)
+                return 0;
+        icmp6h->icmp6_cksum = 0;
+        skb->csum = csum_partial(icmp6h, ipl - ihl, 0);
+        icmp6h->icmp6_cksum = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
+                                              ipl - ihl, IPPROTO_ICMPV6,
+                                              skb->csum);
+        skb->ip_summed = CHECKSUM_NONE;
+        return 1;
+}
+static int tcf_csum_ipv4_tcp(struct sk_buff *skb, struct iphdr *iph,
+                             unsigned int ihl, unsigned int ipl)
+{
+        struct tcphdr *tcph;
+        tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
+        if (tcph == NULL)
+                return 0;
+        tcph->check = 0;
+        skb->csum = csum_partial(tcph, ipl - ihl, 0);
+        tcph->check = tcp_v4_check(ipl - ihl,
+                                   iph->saddr, iph->daddr, skb->csum);
+        skb->ip_summed = CHECKSUM_NONE;
+        return 1;
+}
+static int tcf_csum_ipv6_tcp(struct sk_buff *skb, struct ipv6hdr *ip6h,
+                             unsigned int ihl, unsigned int ipl)
+{
+        struct tcphdr *tcph;
+        tcph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*tcph));
+        if (tcph == NULL)
+                return 0;
+        tcph->check = 0;
+        skb->csum = csum_partial(tcph, ipl - ihl, 0);
+        tcph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr,
+                                      ipl - ihl, IPPROTO_TCP,
+                                      skb->csum);
+        skb->ip_summed = CHECKSUM_NONE;
+        return 1;
+}
+static int tcf_csum_ipv4_udp(struct sk_buff *skb, struct iphdr *iph,
+                             unsigned int ihl, unsigned int ipl, int udplite)
+{
+        struct udphdr *udph;
+        u16 ul;
+        /*
+         * Support both UDP and UDPLITE checksum algorithms, Don't use
+         * udph->len to get the real length without any protocol check,
+         * UDPLITE uses udph->len for another thing,
+         * Use iph->tot_len, or just ipl.
+         */
+        udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
+        if (udph == NULL)
+                return 0;
+        ul = ntohs(udph->len);
+        if (udplite || udph->check) {
+                udph->check = 0;
+                if (udplite) {
+                        if (ul == 0)
+                                skb->csum = csum_partial(udph, ipl - ihl, 0);
+                        else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
+                                skb->csum = csum_partial(udph, ul, 0);
+                        else
+                                goto ignore_obscure_skb;
+                } else {
+                        if (ul != ipl - ihl)
+                                goto ignore_obscure_skb;
+                        skb->csum = csum_partial(udph, ul, 0);
+                }
+                udph->check = csum_tcpudp_magic(iph->saddr, iph->daddr,
+                                                ul, iph->protocol,
+                                                skb->csum);
+                if (!udph->check)
+                        udph->check = CSUM_MANGLED_0;
+        }
+        skb->ip_summed = CHECKSUM_NONE;
+ignore_obscure_skb:
+        return 1;
+}
+static int tcf_csum_ipv6_udp(struct sk_buff *skb, struct ipv6hdr *ip6h,
+                             unsigned int ihl, unsigned int ipl, int udplite)
+{
+        struct udphdr *udph;
+        u16 ul;
+        /*
+         * Support both UDP and UDPLITE checksum algorithms, Don't use
+         * udph->len to get the real length without any protocol check,
+         * UDPLITE uses udph->len for another thing,
+         * Use ip6h->payload_len + sizeof(*ip6h) ... , or just ipl.
+         */
+        udph = tcf_csum_skb_nextlayer(skb, ihl, ipl, sizeof(*udph));
+        if (udph == NULL)
+                return 0;
+        ul = ntohs(udph->len);
+        udph->check = 0;
+        if (udplite) {
+                if (ul == 0)
+                        skb->csum = csum_partial(udph, ipl - ihl, 0);
+                else if ((ul >= sizeof(*udph)) && (ul <= ipl - ihl))
+                        skb->csum = csum_partial(udph, ul, 0);
+                else
+                        goto ignore_obscure_skb;
+        } else {
+                if (ul != ipl - ihl)
+                        goto ignore_obscure_skb;
+                skb->csum = csum_partial(udph, ul, 0);
+        }
+        udph->check = csum_ipv6_magic(&ip6h->saddr, &ip6h->daddr, ul,
+                                      udplite ? IPPROTO_UDPLITE : IPPROTO_UDP,
+                                      skb->csum);
+        if (!udph->check)
+                udph->check = CSUM_MANGLED_0;
+        skb->ip_summed = CHECKSUM_NONE;
+ignore_obscure_skb:
+        return 1;
+}
+static int tcf_csum_ipv4(struct sk_buff *skb, u32 update_flags)
+{
+        struct iphdr *iph;
+        int ntkoff;
+        ntkoff = skb_network_offset(skb);
+        if (!pskb_may_pull(skb, sizeof(*iph) + ntkoff))
+                goto fail;
+        iph = ip_hdr(skb);
+        switch (iph->frag_off & htons(IP_OFFSET) ? 0 : iph->protocol) {
+        case IPPROTO_ICMP:
+                if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
+                        if (!tcf_csum_ipv4_icmp(skb, iph->ihl * 4,
+                                                ntohs(iph->tot_len)))
+                                goto fail;
+                break;
+        case IPPROTO_IGMP:
+                if (update_flags & TCA_CSUM_UPDATE_FLAG_IGMP)
+                        if (!tcf_csum_ipv4_igmp(skb, iph->ihl * 4,
+                                                ntohs(iph->tot_len)))
+                                goto fail;
+                break;
+        case IPPROTO_TCP:
+                if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
+                        if (!tcf_csum_ipv4_tcp(skb, iph, iph->ihl * 4,
+                                               ntohs(iph->tot_len)))
+                                goto fail;
+                break;
+        case IPPROTO_UDP:
+                if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
+                        if (!tcf_csum_ipv4_udp(skb, iph, iph->ihl * 4,
+                                               ntohs(iph->tot_len), 0))
+                                goto fail;
+                break;
+        case IPPROTO_UDPLITE:
+                if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
+                        if (!tcf_csum_ipv4_udp(skb, iph, iph->ihl * 4,
+                                               ntohs(iph->tot_len), 1))
+                                goto fail;
+                break;
+        }
+        if (update_flags & TCA_CSUM_UPDATE_FLAG_IPV4HDR) {
+                if (skb_cloned(skb) &&
+                    !skb_clone_writable(skb, sizeof(*iph) + ntkoff) &&
+                    pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
+                        goto fail;
+                ip_send_check(iph);
+        }
+        return 1;
+fail:
+        return 0;
+}
+static int tcf_csum_ipv6_hopopts(struct ipv6_opt_hdr *ip6xh,
+                                 unsigned int ixhl, unsigned int *pl)
+{
+        int off, len, optlen;
+        unsigned char *xh = (void *)ip6xh;
+        off = sizeof(*ip6xh);
+        len = ixhl - off;
+        while (len > 1) {
+                switch (xh[off]) {
+                case IPV6_TLV_PAD0:
+                        optlen = 1;
+                        break;
+                case IPV6_TLV_JUMBO:
+                        optlen = xh[off + 1] + 2;
+                        if (optlen != 6 || len < 6 || (off & 3) != 2)
+                                /* wrong jumbo option length/alignment */
+                                return 0;
+                        *pl = ntohl(*(__be32 *)(xh + off + 2));
+                        goto done;
+                default:
+                        optlen = xh[off + 1] + 2;
+                        if (optlen > len)
+                                /* ignore obscure options */
+                                goto done;
+                        break;
+                }
+                off += optlen;
+                len -= optlen;
+        }
+done:
+        return 1;
+}
+static int tcf_csum_ipv6(struct sk_buff *skb, u32 update_flags)
+{
+        struct ipv6hdr *ip6h;
+        struct ipv6_opt_hdr *ip6xh;
+        unsigned int hl, ixhl;
+        unsigned int pl;
+        int ntkoff;
+        u8 nexthdr;
+        ntkoff = skb_network_offset(skb);
+        hl = sizeof(*ip6h);
+        if (!pskb_may_pull(skb, hl + ntkoff))
+                goto fail;
+        ip6h = ipv6_hdr(skb);
+        pl = ntohs(ip6h->payload_len);
+        nexthdr = ip6h->nexthdr;
+        do {
+                switch (nexthdr) {
+                case NEXTHDR_FRAGMENT:
+                        goto ignore_skb;
+                case NEXTHDR_ROUTING:
+                case NEXTHDR_HOP:
+                case NEXTHDR_DEST:
+                        if (!pskb_may_pull(skb, hl + sizeof(*ip6xh) + ntkoff))
+                                goto fail;
+                        ip6xh = (void *)(skb_network_header(skb) + hl);
+                        ixhl = ipv6_optlen(ip6xh);
+                        if (!pskb_may_pull(skb, hl + ixhl + ntkoff))
+                                goto fail;
+                        if ((nexthdr == NEXTHDR_HOP) &&
+                            !(tcf_csum_ipv6_hopopts(ip6xh, ixhl, &pl)))
+                                goto fail;
+                        nexthdr = ip6xh->nexthdr;
+                        hl += ixhl;
+                        break;
+                case IPPROTO_ICMPV6:
+                        if (update_flags & TCA_CSUM_UPDATE_FLAG_ICMP)
+                                if (!tcf_csum_ipv6_icmp(skb, ip6h,
+                                                        hl, pl + sizeof(*ip6h)))
+                                        goto fail;
+                        goto done;
+                case IPPROTO_TCP:
+                        if (update_flags & TCA_CSUM_UPDATE_FLAG_TCP)
+                                if (!tcf_csum_ipv6_tcp(skb, ip6h,
+                                                       hl, pl + sizeof(*ip6h)))
+                                        goto fail;
+                        goto done;
+                case IPPROTO_UDP:
+                        if (update_flags & TCA_CSUM_UPDATE_FLAG_UDP)
+                                if (!tcf_csum_ipv6_udp(skb, ip6h, hl,
+                                                       pl + sizeof(*ip6h), 0))
+                                        goto fail;
+                        goto done;
+                case IPPROTO_UDPLITE:
+                        if (update_flags & TCA_CSUM_UPDATE_FLAG_UDPLITE)
+                                if (!tcf_csum_ipv6_udp(skb, ip6h, hl,
+                                                       pl + sizeof(*ip6h), 1))
+                                        goto fail;
+                        goto done;
+                default:
+                        goto ignore_skb;
+                }
+        } while (pskb_may_pull(skb, hl + 1 + ntkoff));
+done:
+ignore_skb:
+        return 1;
+fail:
+        return 0;
+}
+static int tcf_csum(struct sk_buff *skb,
+                    struct tc_action *a, struct tcf_result *res)
+{
+        struct tcf_csum *p = a->priv;
+        int action;
+        u32 update_flags;
+        spin_lock(&p->tcf_lock);
+        p->tcf_tm.lastuse = jiffies;
+        p->tcf_bstats.bytes += qdisc_pkt_len(skb);
+        p->tcf_bstats.packets++;
+        action = p->tcf_action;
+        update_flags = p->update_flags;
+        spin_unlock(&p->tcf_lock);
+        if (unlikely(action == TC_ACT_SHOT))
+                goto drop;
+        switch (skb->protocol) {
+        case cpu_to_be16(ETH_P_IP):
+                if (!tcf_csum_ipv4(skb, update_flags))
+                        goto drop;
+                break;
+        case cpu_to_be16(ETH_P_IPV6):
+                if (!tcf_csum_ipv6(skb, update_flags))
+                        goto drop;
+                break;
+        }
+        return action;
+drop:
+        spin_lock(&p->tcf_lock);
+        p->tcf_qstats.drops++;
+        spin_unlock(&p->tcf_lock);
+        return TC_ACT_SHOT;
+}
+static int tcf_csum_dump(struct sk_buff *skb,
+                         struct tc_action *a, int bind, int ref)
+{
+        unsigned char *b = skb_tail_pointer(skb);
+        struct tcf_csum *p = a->priv;
+        struct tc_csum opt = {
+                .update_flags = p->update_flags,
+                .index   = p->tcf_index,
+                .action  = p->tcf_action,
+                .refcnt  = p->tcf_refcnt - ref,
+                .bindcnt = p->tcf_bindcnt - bind,
+        };
+        struct tcf_t t;
+        NLA_PUT(skb, TCA_CSUM_PARMS, sizeof(opt), &opt);
+        t.install = jiffies_to_clock_t(jiffies - p->tcf_tm.install);
+        t.lastuse = jiffies_to_clock_t(jiffies - p->tcf_tm.lastuse);
+        t.expires = jiffies_to_clock_t(p->tcf_tm.expires);
+        NLA_PUT(skb, TCA_CSUM_TM, sizeof(t), &t);
+        return skb->len;
+nla_put_failure:
+        nlmsg_trim(skb, b);
+        return -1;
+}
+static struct tc_action_ops act_csum_ops = {
+        .kind           = "csum",
+        .hinfo          = &csum_hash_info,
+        .type           = TCA_ACT_CSUM,
+        .capab          = TCA_CAP_NONE,
+        .owner          = THIS_MODULE,
+        .act            = tcf_csum,
+        .dump           = tcf_csum_dump,
+        .cleanup        = tcf_csum_cleanup,
+        .lookup         = tcf_hash_search,
+        .init           = tcf_csum_init,
+        .walk           = tcf_generic_walker
+};
+MODULE_DESCRIPTION("Checksum updating actions");
+MODULE_LICENSE("GPL");
+static int __init csum_init_module(void)
+{
+        return tcf_register_action(&act_csum_ops);
+}
+static void __exit csum_cleanup_module(void)
+{
+        tcf_unregister_action(&act_csum_ops);
+}
+module_init(csum_init_module);
+module_exit(csum_cleanup_module);
diff --git a/net/sched/cls_flow.c b/net/sched/cls_flow.c
index e17096e3913c..5b271a18bc3a 100644
--- a/net/sched/cls_flow.c
+++ b/net/sched/cls_flow.c
@@ -111,44 +111,41 @@ static u32 flow_get_proto(struct sk_buff *skb)
        }
 }
-static int has_ports(u8 protocol)
-{
-        switch (protocol) {
-        case IPPROTO_TCP:
-        case IPPROTO_UDP:
-        case IPPROTO_UDPLITE:
-        case IPPROTO_SCTP:
-        case IPPROTO_DCCP:
-        case IPPROTO_ESP:
-                return 1;
-        default:
-                return 0;
-        }
-}
 static u32 flow_get_proto_src(struct sk_buff *skb)
 {
        switch (skb->protocol) {
        case htons(ETH_P_IP): {
                struct iphdr *iph;
+                int poff;
                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        break;
                iph = ip_hdr(skb);
-                if (!(iph->frag_off&htons(IP_MF|IP_OFFSET)) &&
+                if (iph->frag_off & htons(IP_MF|IP_OFFSET))
-                    has_ports(iph->protocol) &&
+                        break;
-                    pskb_network_may_pull(skb, iph->ihl * 4 + 2))
+                poff = proto_ports_offset(iph->protocol);
-                        return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4));
+                if (poff >= 0 &&
+                    pskb_network_may_pull(skb, iph->ihl * 4 + 2 + poff)) {
+                        iph = ip_hdr(skb);
+                        return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4 +
+                                                 poff));
+                }
                break;
        }
        case htons(ETH_P_IPV6): {
                struct ipv6hdr *iph;
+                int poff;
-                if (!pskb_network_may_pull(skb, sizeof(*iph) + 2))
+                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        break;
                iph = ipv6_hdr(skb);
-                if (has_ports(iph->nexthdr))
+                poff = proto_ports_offset(iph->nexthdr);
-                        return ntohs(*(__be16 *)&iph[1]);
+                if (poff >= 0 &&
+                    pskb_network_may_pull(skb, sizeof(*iph) + poff + 2)) {
+                        iph = ipv6_hdr(skb);
+                        return ntohs(*(__be16 *)((void *)iph + sizeof(*iph) +
+                                                 poff));
+                }
                break;
        }
        }
@@ -161,24 +158,36 @@ static u32 flow_get_proto_dst(struct sk_buff *skb)
        switch (skb->protocol) {
        case htons(ETH_P_IP): {
                struct iphdr *iph;
+                int poff;
                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        break;
                iph = ip_hdr(skb);
-                if (!(iph->frag_off&htons(IP_MF|IP_OFFSET)) &&
+                if (iph->frag_off & htons(IP_MF|IP_OFFSET))
-                    has_ports(iph->protocol) &&
+                        break;
-                    pskb_network_may_pull(skb, iph->ihl * 4 + 4))
+                poff = proto_ports_offset(iph->protocol);
-                        return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4 + 2));
+                if (poff >= 0 &&
+                    pskb_network_may_pull(skb, iph->ihl * 4 + 4 + poff)) {
+                        iph = ip_hdr(skb);
+                        return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4 +
+                                                 2 + poff));
+                }
                break;
        }
        case htons(ETH_P_IPV6): {
                struct ipv6hdr *iph;
+                int poff;
-                if (!pskb_network_may_pull(skb, sizeof(*iph) + 4))
+                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        break;
                iph = ipv6_hdr(skb);
-                if (has_ports(iph->nexthdr))
+                poff = proto_ports_offset(iph->nexthdr);
-                        return ntohs(*(__be16 *)((void *)&iph[1] + 2));
+                if (poff >= 0 &&
+                    pskb_network_may_pull(skb, sizeof(*iph) + poff + 4)) {
+                        iph = ipv6_hdr(skb);
+                        return ntohs(*(__be16 *)((void *)iph + sizeof(*iph) +
+                                                 poff + 2));
+                }
                break;
        }
        }
@@ -297,6 +306,11 @@ static u32 flow_get_vlan_tag(const struct sk_buff *skb)
        return tag & VLAN_VID_MASK;
 }
+static u32 flow_get_rxhash(struct sk_buff *skb)
+{
+        return skb_get_rxhash(skb);
+}
 static u32 flow_key_get(struct sk_buff *skb, int key)
 {
        switch (key) {
@@ -334,6 +348,8 @@ static u32 flow_key_get(struct sk_buff *skb, int key)
                return flow_get_skgid(skb);
        case FLOW_KEY_VLAN_TAG:
                return flow_get_vlan_tag(skb);
+        case FLOW_KEY_RXHASH:
+                return flow_get_rxhash(skb);
        default:
                WARN_ON(1);
                return 0;
diff --git a/net/sched/em_meta.c b/net/sched/em_meta.c
index 3bcac8aa333c..34da5e29ea1a 100644
--- a/net/sched/em_meta.c
+++ b/net/sched/em_meta.c
@@ -223,6 +223,11 @@ META_COLLECTOR(int_maclen)
        dst->value = skb->mac_len;
 }
+META_COLLECTOR(int_rxhash)
+{
+        dst->value = skb_get_rxhash(skb);
+}
 /**************************************************************************
 * Netfilter
 **************************************************************************/
@@ -541,6 +546,7 @@ static struct meta_ops __meta_ops[TCF_META_TYPE_MAX+1][TCF_META_ID_MAX+1] = {
                [META_ID(SK_SENDMSG_OFF)]       = META_FUNC(int_sk_sendmsg_off),
                [META_ID(SK_WRITE_PENDING)]     = META_FUNC(int_sk_write_pend),
                [META_ID(VLAN_TAG)]             = META_FUNC(int_vlan_tag),
+                [META_ID(RXHASH)]               = META_FUNC(int_rxhash),
        }
 };
diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 408eea7086aa..b8020784d0e9 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -240,7 +240,7 @@ struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle)
        if (q)
                goto out;
-        q = qdisc_match_from_root(dev->rx_queue.qdisc_sleeping, handle);
+        q = qdisc_match_from_root(dev->ingress_queue.qdisc_sleeping, handle);
 out:
        return q;
 }
@@ -360,7 +360,7 @@ static struct qdisc_size_table *qdisc_get_stab(struct nlattr *opt)
                tsize = nla_len(tb[TCA_STAB_DATA]) / sizeof(u16);
        }
-        if (!s || tsize != s->tsize || (!tab && tsize > 0))
+        if (tsize != s->tsize || (!tab && tsize > 0))
                return ERR_PTR(-EINVAL);
        spin_lock(&qdisc_stab_lock);
@@ -701,7 +701,7 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent,
                }
                for (i = 0; i < num_q; i++) {
-                        struct netdev_queue *dev_queue = &dev->rx_queue;
+                        struct netdev_queue *dev_queue = &dev->ingress_queue;
                        if (!ingress)
                                dev_queue = netdev_get_tx_queue(dev, i);
@@ -979,7 +979,7 @@ static int tc_get_qdisc(struct sk_buff *skb, struct nlmsghdr *n, void *arg)
                                        return -ENOENT;
                                q = qdisc_leaf(p, clid);
                        } else { /* ingress */
-                                q = dev->rx_queue.qdisc_sleeping;
+                                q = dev->ingress_queue.qdisc_sleeping;
                        }
                } else {
                        q = dev->qdisc;
@@ -1044,7 +1044,7 @@ replay:
                                        return -ENOENT;
                                q = qdisc_leaf(p, clid);
                        } else { /*ingress */
-                                q = dev->rx_queue.qdisc_sleeping;
+                                q = dev->ingress_queue.qdisc_sleeping;
                        }
                } else {
                        q = dev->qdisc;
@@ -1124,7 +1124,7 @@ create_n_graft:
        if (!(n->nlmsg_flags&NLM_F_CREATE))
                return -ENOENT;
        if (clid == TC_H_INGRESS)
-                q = qdisc_create(dev, &dev->rx_queue, p,
+                q = qdisc_create(dev, &dev->ingress_queue, p,
                                 tcm->tcm_parent, tcm->tcm_parent,
                                 tca, &err);
        else {
@@ -1304,7 +1304,7 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb)
                if (tc_dump_qdisc_root(dev->qdisc, skb, cb, &q_idx, s_q_idx) < 0)
                        goto done;
-                dev_queue = &dev->rx_queue;
+                dev_queue = &dev->ingress_queue;
                if (tc_dump_qdisc_root(dev_queue->qdisc_sleeping, skb, cb, &q_idx, s_q_idx) < 0)
                        goto done;
@@ -1595,7 +1595,7 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb)
        if (tc_dump_tclass_root(dev->qdisc, skb, tcm, cb, &t, s_t) < 0)
                goto done;
-        dev_queue = &dev->rx_queue;
+        dev_queue = &dev->ingress_queue;
        if (tc_dump_tclass_root(dev_queue->qdisc_sleeping, skb, tcm, cb, &t, s_t) < 0)
                goto done;
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 2aeb3a4386a1..545278a1c478 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -753,7 +753,7 @@ void dev_activate(struct net_device *dev)
        need_watchdog = 0;
        netdev_for_each_tx_queue(dev, transition_one_qdisc, &need_watchdog);
-        transition_one_qdisc(dev, &dev->rx_queue, NULL);
+        transition_one_qdisc(dev, &dev->ingress_queue, NULL);
        if (need_watchdog) {
                dev->trans_start = jiffies;
@@ -812,7 +812,7 @@ static bool some_qdisc_is_busy(struct net_device *dev)
 void dev_deactivate(struct net_device *dev)
 {
        netdev_for_each_tx_queue(dev, dev_deactivate_queue, &noop_qdisc);
-        dev_deactivate_queue(dev, &dev->rx_queue, &noop_qdisc);
+        dev_deactivate_queue(dev, &dev->ingress_queue, &noop_qdisc);
        dev_watchdog_down(dev);
@@ -838,7 +838,7 @@ void dev_init_scheduler(struct net_device *dev)
 {
        dev->qdisc = &noop_qdisc;
        netdev_for_each_tx_queue(dev, dev_init_scheduler_queue, &noop_qdisc);
-        dev_init_scheduler_queue(dev, &dev->rx_queue, &noop_qdisc);
+        dev_init_scheduler_queue(dev, &dev->ingress_queue, &noop_qdisc);
        setup_timer(&dev->watchdog_timer, dev_watchdog, (unsigned long)dev);
 }
@@ -861,7 +861,7 @@ static void shutdown_scheduler_queue(struct net_device *dev,
 void dev_shutdown(struct net_device *dev)
 {
        netdev_for_each_tx_queue(dev, shutdown_scheduler_queue, &noop_qdisc);
-        shutdown_scheduler_queue(dev, &dev->rx_queue, &noop_qdisc);
+        shutdown_scheduler_queue(dev, &dev->ingress_queue, &noop_qdisc);
        qdisc_destroy(dev->qdisc);
        dev->qdisc = &noop_qdisc;
diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c
index 201cbac2b32c..3cf478d012dd 100644
--- a/net/sched/sch_sfq.c
+++ b/net/sched/sch_sfq.c
@@ -123,40 +123,39 @@ static unsigned sfq_hash(struct sfq_sched_data *q, struct sk_buff *skb)
        case htons(ETH_P_IP):
        {
                const struct iphdr *iph;
+                int poff;
                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        goto err;
                iph = ip_hdr(skb);
                h = (__force u32)iph->daddr;
                h2 = (__force u32)iph->saddr ^ iph->protocol;
-                if (!(iph->frag_off&htons(IP_MF|IP_OFFSET)) &&
+                if (iph->frag_off & htons(IP_MF|IP_OFFSET))
-                    (iph->protocol == IPPROTO_TCP ||
+                        break;
-                     iph->protocol == IPPROTO_UDP ||
+                poff = proto_ports_offset(iph->protocol);
-                     iph->protocol == IPPROTO_UDPLITE ||
+                if (poff >= 0 &&
-                     iph->protocol == IPPROTO_SCTP ||
+                    pskb_network_may_pull(skb, iph->ihl * 4 + 4 + poff)) {
-                     iph->protocol == IPPROTO_DCCP ||
+                        iph = ip_hdr(skb);
-                     iph->protocol == IPPROTO_ESP) &&
+                        h2 ^= *(u32*)((void *)iph + iph->ihl * 4 + poff);
-                     pskb_network_may_pull(skb, iph->ihl * 4 + 4))
+                }
-                        h2 ^= *(((u32*)iph) + iph->ihl);
                break;
        }
        case htons(ETH_P_IPV6):
        {
                struct ipv6hdr *iph;
+                int poff;
                if (!pskb_network_may_pull(skb, sizeof(*iph)))
                        goto err;
                iph = ipv6_hdr(skb);
                h = (__force u32)iph->daddr.s6_addr32[3];
                h2 = (__force u32)iph->saddr.s6_addr32[3] ^ iph->nexthdr;
-                if ((iph->nexthdr == IPPROTO_TCP ||
+                poff = proto_ports_offset(iph->nexthdr);
-                     iph->nexthdr == IPPROTO_UDP ||
+                if (poff >= 0 &&
-                     iph->nexthdr == IPPROTO_UDPLITE ||
+                    pskb_network_may_pull(skb, sizeof(*iph) + 4 + poff)) {
-                     iph->nexthdr == IPPROTO_SCTP ||
+                        iph = ipv6_hdr(skb);
-                     iph->nexthdr == IPPROTO_DCCP ||
+                        h2 ^= *(u32*)((void *)iph + sizeof(*iph) + poff);
-                     iph->nexthdr == IPPROTO_ESP) &&
+                }
-                    pskb_network_may_pull(skb, sizeof(*iph) + 4))
-                        h2 ^= *(u32*)&iph[1];
                break;
        }
        default:
diff --git a/net/sctp/associola.c b/net/sctp/associola.c
index 0b85e5256434..5f1fb8bd862d 100644
--- a/net/sctp/associola.c
+++ b/net/sctp/associola.c
@@ -48,6 +48,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/fcntl.h>
 #include <linux/poll.h>
diff --git a/net/sctp/chunk.c b/net/sctp/chunk.c
index 476caaf100ed..6c8556459a75 100644
--- a/net/sctp/chunk.c
+++ b/net/sctp/chunk.c
@@ -37,6 +37,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/net.h>
diff --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c
index ccb6dc48d15b..397296fb156f 100644
--- a/net/sctp/inqueue.c
+++ b/net/sctp/inqueue.c
@@ -43,6 +43,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <net/sctp/sctp.h>
 #include <net/sctp/sm.h>
 #include <linux/interrupt.h>
diff --git a/net/sctp/ipv6.c b/net/sctp/ipv6.c
index 732689140fb8..95e0c8eda1a0 100644
--- a/net/sctp/ipv6.c
+++ b/net/sctp/ipv6.c
@@ -47,6 +47,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/module.h>
 #include <linux/errno.h>
 #include <linux/types.h>
@@ -336,7 +338,7 @@ static void sctp_v6_get_saddr(struct sctp_sock *sk,
                memcpy(saddr, baddr, sizeof(union sctp_addr));
                SCTP_DEBUG_PRINTK("saddr: %pI6\n", &saddr->v6.sin6_addr);
        } else {
-                printk(KERN_ERR "%s: asoc:%p Could not find a valid source "
+                pr_err("%s: asoc:%p Could not find a valid source "
                       "address for the dest:%pI6\n",
                       __func__, asoc, &daddr->v6.sin6_addr);
        }
diff --git a/net/sctp/objcnt.c b/net/sctp/objcnt.c
index f73ec0ea93ba..8ef8e7d9eb61 100644
--- a/net/sctp/objcnt.c
+++ b/net/sctp/objcnt.c
@@ -38,6 +38,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/kernel.h>
 #include <net/sctp/sctp.h>
@@ -134,8 +136,7 @@ void sctp_dbg_objcnt_init(void)
        ent = proc_create("sctp_dbg_objcnt", 0,
                          proc_net_sctp, &sctp_objcnt_ops);
        if (!ent)
-                printk(KERN_WARNING
+                pr_warn("sctp_dbg_objcnt: Unable to create /proc entry.\n");
-                        "sctp_dbg_objcnt: Unable to create /proc entry.\n");
 }
 /* Cleanup the objcount entry in the proc filesystem.  */
diff --git a/net/sctp/output.c b/net/sctp/output.c
index bcc4590ccaf2..60600d337a3a 100644
--- a/net/sctp/output.c
+++ b/net/sctp/output.c
@@ -41,6 +41,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/wait.h>
diff --git a/net/sctp/outqueue.c b/net/sctp/outqueue.c
index c04b2eb59186..8c6d379b4bb6 100644
--- a/net/sctp/outqueue.c
+++ b/net/sctp/outqueue.c
@@ -46,6 +46,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/list.h>   /* For struct list_head */
 #include <linux/socket.h>
@@ -1463,23 +1465,23 @@ static void sctp_check_transmitted(struct sctp_outq *q,
                                        /* Display the end of the
                                         * current range.
                                         */
-                                        SCTP_DEBUG_PRINTK("-%08x",
+                                        SCTP_DEBUG_PRINTK_CONT("-%08x",
-                                                          dbg_last_ack_tsn);
+                                                               dbg_last_ack_tsn);
                                }
                                /* Start a new range.  */
-                                SCTP_DEBUG_PRINTK(",%08x", tsn);
+                                SCTP_DEBUG_PRINTK_CONT(",%08x", tsn);
                                dbg_ack_tsn = tsn;
                                break;
                        case 1: /* The last TSN was NOT ACKed. */
                                if (dbg_last_kept_tsn != dbg_kept_tsn) {
                                        /* Display the end of current range. */
-                                        SCTP_DEBUG_PRINTK("-%08x",
+                                        SCTP_DEBUG_PRINTK_CONT("-%08x",
-                                                          dbg_last_kept_tsn);
+                                                               dbg_last_kept_tsn);
                                }
-                                SCTP_DEBUG_PRINTK("\n");
+                                SCTP_DEBUG_PRINTK_CONT("\n");
                                /* FALL THROUGH... */
                        default:
@@ -1526,18 +1528,18 @@ static void sctp_check_transmitted(struct sctp_outq *q,
                                        break;
                                if (dbg_last_kept_tsn != dbg_kept_tsn)
-                                        SCTP_DEBUG_PRINTK("-%08x",
+                                        SCTP_DEBUG_PRINTK_CONT("-%08x",
-                                                          dbg_last_kept_tsn);
+                                                               dbg_last_kept_tsn);
-                                SCTP_DEBUG_PRINTK(",%08x", tsn);
+                                SCTP_DEBUG_PRINTK_CONT(",%08x", tsn);
                                dbg_kept_tsn = tsn;
                                break;
                        case 0:
                                if (dbg_last_ack_tsn != dbg_ack_tsn)
-                                        SCTP_DEBUG_PRINTK("-%08x",
+                                        SCTP_DEBUG_PRINTK_CONT("-%08x",
-                                                          dbg_last_ack_tsn);
+                                                               dbg_last_ack_tsn);
-                                SCTP_DEBUG_PRINTK("\n");
+                                SCTP_DEBUG_PRINTK_CONT("\n");
                                /* FALL THROUGH... */
                        default:
@@ -1556,17 +1558,17 @@ static void sctp_check_transmitted(struct sctp_outq *q,
        switch (dbg_prt_state) {
        case 0:
                if (dbg_last_ack_tsn != dbg_ack_tsn) {
-                        SCTP_DEBUG_PRINTK("-%08x\n", dbg_last_ack_tsn);
+                        SCTP_DEBUG_PRINTK_CONT("-%08x\n", dbg_last_ack_tsn);
                } else {
-                        SCTP_DEBUG_PRINTK("\n");
+                        SCTP_DEBUG_PRINTK_CONT("\n");
                }
        break;
        case 1:
                if (dbg_last_kept_tsn != dbg_kept_tsn) {
-                        SCTP_DEBUG_PRINTK("-%08x\n", dbg_last_kept_tsn);
+                        SCTP_DEBUG_PRINTK_CONT("-%08x\n", dbg_last_kept_tsn);
                } else {
-                        SCTP_DEBUG_PRINTK("\n");
+                        SCTP_DEBUG_PRINTK_CONT("\n");
                }
        }
 #endif /* SCTP_DEBUG */
diff --git a/net/sctp/probe.c b/net/sctp/probe.c
index db3a42b8b349..2e63e9dc010e 100644
--- a/net/sctp/probe.c
+++ b/net/sctp/probe.c
@@ -22,6 +22,8 @@
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/kernel.h>
 #include <linux/kprobes.h>
 #include <linux/socket.h>
@@ -192,7 +194,7 @@ static __init int sctpprobe_init(void)
        if (ret)
                goto remove_proc;
-        pr_info("SCTP probe registered (port=%d)\n", port);
+        pr_info("probe registered (port=%d)\n", port);
        return 0;
diff --git a/net/sctp/protocol.c b/net/sctp/protocol.c
index 5027b83f1cc0..1ef29c74d85e 100644
--- a/net/sctp/protocol.c
+++ b/net/sctp/protocol.c
@@ -46,6 +46,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/netdevice.h>
@@ -707,8 +709,7 @@ static int sctp_ctl_sock_init(void)
                                           &init_net);
        if (err < 0) {
-                printk(KERN_ERR
+                pr_err("Failed to create the SCTP control socket\n");
-                       "SCTP: Failed to create the SCTP control socket.\n");
                return err;
        }
        return 0;
@@ -798,7 +799,7 @@ static void sctp_inet_skb_msgname(struct sk_buff *skb, char *msgname, int *len)
 static int sctp_inet_af_supported(sa_family_t family, struct sctp_sock *sp)
 {
        /* PF_INET only supports AF_INET addresses. */
-        return (AF_INET == family);
+        return AF_INET == family;
 }
 /* Address matching with wildcards allowed. */
@@ -1206,7 +1207,7 @@ SCTP_STATIC __init int sctp_init(void)
                                        __get_free_pages(GFP_ATOMIC, order);
        } while (!sctp_assoc_hashtable && --order > 0);
        if (!sctp_assoc_hashtable) {
-                printk(KERN_ERR "SCTP: Failed association hash alloc.\n");
+                pr_err("Failed association hash alloc\n");
                status = -ENOMEM;
                goto err_ahash_alloc;
        }
@@ -1220,7 +1221,7 @@ SCTP_STATIC __init int sctp_init(void)
        sctp_ep_hashtable = (struct sctp_hashbucket *)
                kmalloc(64 * sizeof(struct sctp_hashbucket), GFP_KERNEL);
        if (!sctp_ep_hashtable) {
-                printk(KERN_ERR "SCTP: Failed endpoint_hash alloc.\n");
+                pr_err("Failed endpoint_hash alloc\n");
                status = -ENOMEM;
                goto err_ehash_alloc;
        }
@@ -1239,7 +1240,7 @@ SCTP_STATIC __init int sctp_init(void)
                                        __get_free_pages(GFP_ATOMIC, order);
        } while (!sctp_port_hashtable && --order > 0);
        if (!sctp_port_hashtable) {
-                printk(KERN_ERR "SCTP: Failed bind hash alloc.");
+                pr_err("Failed bind hash alloc\n");
                status = -ENOMEM;
                goto err_bhash_alloc;
        }
@@ -1248,8 +1249,7 @@ SCTP_STATIC __init int sctp_init(void)
                INIT_HLIST_HEAD(&sctp_port_hashtable[i].chain);
        }
-        printk(KERN_INFO "SCTP: Hash tables configured "
+        pr_info("Hash tables configured (established %d bind %d)\n",
-                         "(established %d bind %d)\n",
                sctp_assoc_hashsize, sctp_port_hashsize);
        /* Disable ADDIP by default. */
@@ -1290,8 +1290,7 @@ SCTP_STATIC __init int sctp_init(void)
        /* Initialize the control inode/socket for handling OOTB packets.  */
        if ((status = sctp_ctl_sock_init())) {
-                printk (KERN_ERR
+                pr_err("Failed to initialize the SCTP control sock\n");
-                        "SCTP: Failed to initialize the SCTP control sock.\n");
                goto err_ctl_sock_init;
        }
diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index 246f92924658..2cc46f0962ca 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -50,6 +50,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/ip.h>
diff --git a/net/sctp/sm_sideeffect.c b/net/sctp/sm_sideeffect.c
index f5e5e27cac5e..b21b218d564f 100644
--- a/net/sctp/sm_sideeffect.c
+++ b/net/sctp/sm_sideeffect.c
@@ -47,6 +47,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/skbuff.h>
 #include <linux/types.h>
 #include <linux/socket.h>
@@ -1146,26 +1148,23 @@ static int sctp_side_effects(sctp_event_t event_type, sctp_subtype_t subtype,
        case SCTP_DISPOSITION_VIOLATION:
                if (net_ratelimit())
-                        printk(KERN_ERR "sctp protocol violation state %d "
+                        pr_err("protocol violation state %d chunkid %d\n",
-                               "chunkid %d\n", state, subtype.chunk);
+                               state, subtype.chunk);
                break;
        case SCTP_DISPOSITION_NOT_IMPL:
-                printk(KERN_WARNING "sctp unimplemented feature in state %d, "
+                pr_warn("unimplemented feature in state %d, event_type %d, event_id %d\n",
-                       "event_type %d, event_id %d\n",
+                        state, event_type, subtype.chunk);
-                       state, event_type, subtype.chunk);
                break;
        case SCTP_DISPOSITION_BUG:
-                printk(KERN_ERR "sctp bug in state %d, "
+                pr_err("bug in state %d, event_type %d, event_id %d\n",
-                       "event_type %d, event_id %d\n",
                       state, event_type, subtype.chunk);
                BUG();
                break;
        default:
-                printk(KERN_ERR "sctp impossible disposition %d "
+                pr_err("impossible disposition %d in state %d, event_type %d, event_id %d\n",
-                       "in state %d, event_type %d, event_id %d\n",
                       status, state, event_type, subtype.chunk);
                BUG();
                break;
@@ -1679,8 +1678,8 @@ static int sctp_cmd_interpreter(sctp_event_t event_type,
                        sctp_cmd_send_asconf(asoc);
                        break;
                default:
-                        printk(KERN_WARNING "Impossible command: %u, %p\n",
+                        pr_warn("Impossible command: %u, %p\n",
-                               cmd->verb, cmd->obj.ptr);
+                                cmd->verb, cmd->obj.ptr);
                        break;
                }
diff --git a/net/sctp/sm_statefuns.c b/net/sctp/sm_statefuns.c
index d344dc481ccc..4b4eb7c96bbd 100644
--- a/net/sctp/sm_statefuns.c
+++ b/net/sctp/sm_statefuns.c
@@ -50,6 +50,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/ip.h>
@@ -1138,18 +1140,16 @@ sctp_disposition_t sctp_sf_backbeat_8_3(const struct sctp_endpoint *ep,
        if (unlikely(!link)) {
                if (from_addr.sa.sa_family == AF_INET6) {
                        if (net_ratelimit())
-                                printk(KERN_WARNING
+                                pr_warn("%s association %p could not find address %pI6\n",
-                                    "%s association %p could not find address %pI6\n",
+                                        __func__,
-                                    __func__,
+                                        asoc,
-                                    asoc,
+                                        &from_addr.v6.sin6_addr);
-                                    &from_addr.v6.sin6_addr);
                } else {
                        if (net_ratelimit())
-                                printk(KERN_WARNING
+                                pr_warn("%s association %p could not find address %pI4\n",
-                                    "%s association %p could not find address %pI4\n",
+                                        __func__,
-                                    __func__,
+                                        asoc,
-                                    asoc,
+                                        &from_addr.v4.sin_addr.s_addr);
-                                    &from_addr.v4.sin_addr.s_addr);
                }
                return SCTP_DISPOSITION_DISCARD;
        }
diff --git a/net/sctp/sm_statetable.c b/net/sctp/sm_statetable.c
index 6d9b3aafcc5d..546d4387fb3c 100644
--- a/net/sctp/sm_statetable.c
+++ b/net/sctp/sm_statetable.c
@@ -46,6 +46,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/skbuff.h>
 #include <net/sctp/sctp.h>
 #include <net/sctp/sm.h>
@@ -66,15 +68,19 @@ static const sctp_sm_table_entry_t bug = {
        .name = "sctp_sf_bug"
 };
-#define DO_LOOKUP(_max, _type, _table) \
+#define DO_LOOKUP(_max, _type, _table)                                  \
-        if ((event_subtype._type > (_max))) { \
+({                                                                      \
-                printk(KERN_WARNING \
+        const sctp_sm_table_entry_t *rtn;                               \
-                       "sctp table %p possible attack:" \
+                                                                        \
-                       " event %d exceeds max %d\n", \
+        if ((event_subtype._type > (_max))) {                           \
-                       _table, event_subtype._type, _max); \
+                pr_warn("table %p possible attack: event %d exceeds max %d\n", \
-                return &bug; \
+                        _table, event_subtype._type, _max);             \
-        } \
+                rtn = &bug;                                             \
-        return &_table[event_subtype._type][(int)state];
+        } else                                                          \
+                rtn = &_table[event_subtype._type][(int)state];         \
+                                                                        \
+        rtn;                                                            \
+})
 const sctp_sm_table_entry_t *sctp_sm_lookup_event(sctp_event_t event_type,
                                                  sctp_state_t state,
@@ -83,21 +89,15 @@ const sctp_sm_table_entry_t *sctp_sm_lookup_event(sctp_event_t event_type,
        switch (event_type) {
        case SCTP_EVENT_T_CHUNK:
                return sctp_chunk_event_lookup(event_subtype.chunk, state);
-                break;
        case SCTP_EVENT_T_TIMEOUT:
-                DO_LOOKUP(SCTP_EVENT_TIMEOUT_MAX, timeout,
+                return DO_LOOKUP(SCTP_EVENT_TIMEOUT_MAX, timeout,
-                          timeout_event_table);
+                                 timeout_event_table);
-                break;
        case SCTP_EVENT_T_OTHER:
-                DO_LOOKUP(SCTP_EVENT_OTHER_MAX, other, other_event_table);
+                return DO_LOOKUP(SCTP_EVENT_OTHER_MAX, other,
-                break;
+                                 other_event_table);
        case SCTP_EVENT_T_PRIMITIVE:
-                DO_LOOKUP(SCTP_EVENT_PRIMITIVE_MAX, primitive,
+                return DO_LOOKUP(SCTP_EVENT_PRIMITIVE_MAX, primitive,
-                          primitive_event_table);
+                                 primitive_event_table);
-                break;
        default:
                /* Yikes!  We got an illegal event type.  */
                return &bug;
diff --git a/net/sctp/socket.c b/net/sctp/socket.c
index fbb70770ad05..e34ca9cc1167 100644
--- a/net/sctp/socket.c
+++ b/net/sctp/socket.c
@@ -57,6 +57,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/types.h>
 #include <linux/kernel.h>
 #include <linux/wait.h>
@@ -2469,9 +2471,8 @@ static int sctp_setsockopt_delayed_ack(struct sock *sk,
                if (params.sack_delay == 0 && params.sack_freq == 0)
                        return 0;
        } else if (optlen == sizeof(struct sctp_assoc_value)) {
-                printk(KERN_WARNING "SCTP: Use of struct sctp_assoc_value "
+                pr_warn("Use of struct sctp_assoc_value in delayed_ack socket option deprecated\n");
-                       "in delayed_ack socket option deprecated\n");
+                pr_warn("Use struct sctp_sack_info instead\n");
-                printk(KERN_WARNING "SCTP: Use struct sctp_sack_info instead\n");
                if (copy_from_user(&params, optval, optlen))
                        return -EFAULT;
@@ -2879,10 +2880,8 @@ static int sctp_setsockopt_maxseg(struct sock *sk, char __user *optval, unsigned
        int val;
        if (optlen == sizeof(int)) {
-                printk(KERN_WARNING
+                pr_warn("Use of int in maxseg socket option deprecated\n");
-                   "SCTP: Use of int in maxseg socket option deprecated\n");
+                pr_warn("Use struct sctp_assoc_value instead\n");
-                printk(KERN_WARNING
-                   "SCTP: Use struct sctp_assoc_value instead\n");
                if (copy_from_user(&val, optval, optlen))
                        return -EFAULT;
                params.assoc_id = 0;
@@ -3132,10 +3131,8 @@ static int sctp_setsockopt_maxburst(struct sock *sk,
        int assoc_id = 0;
        if (optlen == sizeof(int)) {
-                printk(KERN_WARNING
+                pr_warn("Use of int in max_burst socket option deprecated\n");
-                   "SCTP: Use of int in max_burst socket option deprecated\n");
+                pr_warn("Use struct sctp_assoc_value instead\n");
-                printk(KERN_WARNING
-                   "SCTP: Use struct sctp_assoc_value instead\n");
                if (copy_from_user(&val, optval, optlen))
                        return -EFAULT;
        } else if (optlen == sizeof(struct sctp_assoc_value)) {
@@ -3606,7 +3603,40 @@ out:
 /* The SCTP ioctl handler. */
 SCTP_STATIC int sctp_ioctl(struct sock *sk, int cmd, unsigned long arg)
 {
-        return -ENOIOCTLCMD;
+        int rc = -ENOTCONN;
+        sctp_lock_sock(sk);
+        /*
+         * SEQPACKET-style sockets in LISTENING state are valid, for
+         * SCTP, so only discard TCP-style sockets in LISTENING state.
+         */
+        if (sctp_style(sk, TCP) && sctp_sstate(sk, LISTENING))
+                goto out;
+        switch (cmd) {
+        case SIOCINQ: {
+                struct sk_buff *skb;
+                unsigned int amount = 0;
+                skb = skb_peek(&sk->sk_receive_queue);
+                if (skb != NULL) {
+                        /*
+                         * We will only return the amount of this packet since
+                         * that is all that will be read.
+                         */
+                        amount = skb->len;
+                }
+                rc = put_user(amount, (int __user *)arg);
+                break;
+        }
+        default:
+                rc = -ENOIOCTLCMD;
+                break;
+        }
+out:
+        sctp_release_sock(sk);
+        return rc;
 }
 /* This is the function which gets called during socket creation to
@@ -3865,7 +3895,7 @@ static int sctp_getsockopt_sctp_status(struct sock *sk, int len,
        }
 out:
-        return (retval);
+        return retval;
 }
@@ -3921,7 +3951,7 @@ static int sctp_getsockopt_peer_addr_info(struct sock *sk, int len,
        }
 out:
-        return (retval);
+        return retval;
 }
 /* 7.1.12 Enable/Disable message fragmentation (SCTP_DISABLE_FRAGMENTS)
@@ -4292,9 +4322,8 @@ static int sctp_getsockopt_delayed_ack(struct sock *sk, int len,
                if (copy_from_user(&params, optval, len))
                        return -EFAULT;
        } else if (len == sizeof(struct sctp_assoc_value)) {
-                printk(KERN_WARNING "SCTP: Use of struct sctp_assoc_value "
+                pr_warn("Use of struct sctp_assoc_value in delayed_ack socket option deprecated\n");
-                       "in delayed_ack socket option deprecated\n");
+                pr_warn("Use struct sctp_sack_info instead\n");
-                printk(KERN_WARNING "SCTP: Use struct sctp_sack_info instead\n");
                if (copy_from_user(&params, optval, len))
                        return -EFAULT;
        } else
@@ -4940,10 +4969,8 @@ static int sctp_getsockopt_maxseg(struct sock *sk, int len,
        struct sctp_association *asoc;
        if (len == sizeof(int)) {
-                printk(KERN_WARNING
+                pr_warn("Use of int in maxseg socket option deprecated\n");
-                   "SCTP: Use of int in maxseg socket option deprecated\n");
+                pr_warn("Use struct sctp_assoc_value instead\n");
-                printk(KERN_WARNING
-                   "SCTP: Use struct sctp_assoc_value instead\n");
                params.assoc_id = 0;
        } else if (len >= sizeof(struct sctp_assoc_value)) {
                len = sizeof(struct sctp_assoc_value);
@@ -5034,10 +5061,8 @@ static int sctp_getsockopt_maxburst(struct sock *sk, int len,
        struct sctp_association *asoc;
        if (len == sizeof(int)) {
-                printk(KERN_WARNING
+                pr_warn("Use of int in max_burst socket option deprecated\n");
-                   "SCTP: Use of int in max_burst socket option deprecated\n");
+                pr_warn("Use struct sctp_assoc_value instead\n");
-                printk(KERN_WARNING
-                   "SCTP: Use struct sctp_assoc_value instead\n");
                params.assoc_id = 0;
        } else if (len >= sizeof(struct sctp_assoc_value)) {
                len = sizeof(struct sctp_assoc_value);
@@ -5580,7 +5605,7 @@ static int sctp_get_port(struct sock *sk, unsigned short snum)
        /* Note: sk->sk_num gets filled in if ephemeral port request. */
        ret = sctp_get_port_local(sk, &addr);
-        return (ret ? 1 : 0);
+        return ret ? 1 : 0;
 }
 /*
@@ -5597,8 +5622,7 @@ SCTP_STATIC int sctp_listen_start(struct sock *sk, int backlog)
                tfm = crypto_alloc_hash(sctp_hmac_alg, 0, CRYPTO_ALG_ASYNC);
                if (IS_ERR(tfm)) {
                        if (net_ratelimit()) {
-                                printk(KERN_INFO
+                                pr_info("failed to load transform for %s: %ld\n",
-                                       "SCTP: failed to load transform for %s: %ld\n",
                                        sctp_hmac_alg, PTR_ERR(tfm));
                        }
                        return -ENOSYS;
@@ -5727,13 +5751,12 @@ unsigned int sctp_poll(struct file *file, struct socket *sock, poll_table *wait)
        if (sk->sk_err || !skb_queue_empty(&sk->sk_error_queue))
                mask |= POLLERR;
        if (sk->sk_shutdown & RCV_SHUTDOWN)
-                mask |= POLLRDHUP;
+                mask |= POLLRDHUP | POLLIN | POLLRDNORM;
        if (sk->sk_shutdown == SHUTDOWN_MASK)
                mask |= POLLHUP;
        /* Is it readable?  Reconsider this code with TCP-style support.  */
-        if (!skb_queue_empty(&sk->sk_receive_queue) ||
+        if (!skb_queue_empty(&sk->sk_receive_queue))
-            (sk->sk_shutdown & RCV_SHUTDOWN))
                mask |= POLLIN | POLLRDNORM;
        /* The association is either gone or not ready.  */
diff --git a/net/sctp/transport.c b/net/sctp/transport.c
index 132046cb82fc..d3ae493d234a 100644
--- a/net/sctp/transport.c
+++ b/net/sctp/transport.c
@@ -48,6 +48,8 @@
 * be incorporated into the next SCTP release.
 */
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
 #include <linux/slab.h>
 #include <linux/types.h>
 #include <linux/random.h>
@@ -244,10 +246,9 @@ void sctp_transport_update_pmtu(struct sctp_transport *t, u32 pmtu)
        struct dst_entry *dst;
        if (unlikely(pmtu < SCTP_DEFAULT_MINSEGMENT)) {
-                printk(KERN_WARNING "%s: Reported pmtu %d too low, "
+                pr_warn("%s: Reported pmtu %d too low, using default minimum of %d\n",
-                       "using default minimum of %d\n",
+                        __func__, pmtu,
-                       __func__, pmtu,
+                        SCTP_DEFAULT_MINSEGMENT);
-                       SCTP_DEFAULT_MINSEGMENT);
                /* Use default minimum segment size and disable
                 * pmtu discovery on this transport.
                 */
diff --git a/net/socket.c b/net/socket.c
index 2270b941bcc7..717a5f1c8792 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -535,14 +535,13 @@ void sock_release(struct socket *sock)
 }
 EXPORT_SYMBOL(sock_release);
-int sock_tx_timestamp(struct msghdr *msg, struct sock *sk,
+int sock_tx_timestamp(struct sock *sk, __u8 *tx_flags)
-                      union skb_shared_tx *shtx)
 {
-        shtx->flags = 0;
+        *tx_flags = 0;
        if (sock_flag(sk, SOCK_TIMESTAMPING_TX_HARDWARE))
-                shtx->hardware = 1;
+                *tx_flags |= SKBTX_HW_TSTAMP;
        if (sock_flag(sk, SOCK_TIMESTAMPING_TX_SOFTWARE))
-                shtx->software = 1;
+                *tx_flags |= SKBTX_SW_TSTAMP;
        return 0;
 }
 EXPORT_SYMBOL(sock_tx_timestamp);
@@ -1919,7 +1918,8 @@ SYSCALL_DEFINE3(sendmsg, int, fd, struct msghdr __user *, msg, unsigned, flags)
                 * Afterwards, it will be a kernel pointer. Thus the compiler-assisted
                 * checking falls down on this.
                 */
-                if (copy_from_user(ctl_buf, (void __user *)msg_sys.msg_control,
+                if (copy_from_user(ctl_buf,
+                                   (void __user __force *)msg_sys.msg_control,
                                   ctl_len))
                        goto out_freectl;
                msg_sys.msg_control = ctl_buf;
@@ -3054,14 +3054,19 @@ int kernel_getsockopt(struct socket *sock, int level, int optname,
                        char *optval, int *optlen)
 {
        mm_segment_t oldfs = get_fs();
+        char __user *uoptval;
+        int __user *uoptlen;
        int err;
+        uoptval = (char __user __force *) optval;
+        uoptlen = (int __user __force *) optlen;
        set_fs(KERNEL_DS);
        if (level == SOL_SOCKET)
-                err = sock_getsockopt(sock, level, optname, optval, optlen);
+                err = sock_getsockopt(sock, level, optname, uoptval, uoptlen);
        else
-                err = sock->ops->getsockopt(sock, level, optname, optval,
+                err = sock->ops->getsockopt(sock, level, optname, uoptval,
-                                            optlen);
+                                            uoptlen);
        set_fs(oldfs);
        return err;
 }
@@ -3071,13 +3076,16 @@ int kernel_setsockopt(struct socket *sock, int level, int optname,
                        char *optval, unsigned int optlen)
 {
        mm_segment_t oldfs = get_fs();
+        char __user *uoptval;
        int err;
+        uoptval = (char __user __force *) optval;
        set_fs(KERNEL_DS);
        if (level == SOL_SOCKET)
-                err = sock_setsockopt(sock, level, optname, optval, optlen);
+                err = sock_setsockopt(sock, level, optname, uoptval, optlen);
        else
-                err = sock->ops->setsockopt(sock, level, optname, optval,
+                err = sock->ops->setsockopt(sock, level, optname, uoptval,
                                            optlen);
        set_fs(oldfs);
        return err;
diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c
index dcfc66bab2bb..597c493392ad 100644
--- a/net/sunrpc/auth_gss/auth_gss.c
+++ b/net/sunrpc/auth_gss/auth_gss.c
@@ -1049,7 +1049,7 @@ gss_match(struct auth_cred *acred, struct rpc_cred *rc, int flags)
 out:
        if (acred->machine_cred != gss_cred->gc_machine_cred)
                return 0;
-        return (rc->cr_uid == acred->uid);
+        return rc->cr_uid == acred->uid;
 }
 /*
diff --git a/net/sunrpc/auth_gss/gss_generic_token.c b/net/sunrpc/auth_gss/gss_generic_token.c
index 310b78e99456..c586e92bcf76 100644
--- a/net/sunrpc/auth_gss/gss_generic_token.c
+++ b/net/sunrpc/auth_gss/gss_generic_token.c
@@ -76,19 +76,19 @@ static int
 der_length_size( int length)
 {
        if (length < (1<<7))
-                return(1);
+                return 1;
        else if (length < (1<<8))
-                return(2);
+                return 2;
 #if (SIZEOF_INT == 2)
        else
-                return(3);
+                return 3;
 #else
        else if (length < (1<<16))
-                return(3);
+                return 3;
        else if (length < (1<<24))
-                return(4);
+                return 4;
        else
-                return(5);
+                return 5;
 #endif
 }
@@ -121,14 +121,14 @@ der_read_length(unsigned char **buf, int *bufsize)
        int ret;
        if (*bufsize < 1)
-                return(-1);
+                return -1;
        sf = *(*buf)++;
        (*bufsize)--;
        if (sf & 0x80) {
                if ((sf &= 0x7f) > ((*bufsize)-1))
-                        return(-1);
+                        return -1;
                if (sf > SIZEOF_INT)
-                        return (-1);
+                        return -1;
                ret = 0;
                for (; sf; sf--) {
                        ret = (ret<<8) + (*(*buf)++);
@@ -138,7 +138,7 @@ der_read_length(unsigned char **buf, int *bufsize)
                ret = sf;
        }
-        return(ret);
+        return ret;
 }
 /* returns the length of a token, given the mech oid and the body size */
@@ -148,7 +148,7 @@ g_token_size(struct xdr_netobj *mech, unsigned int body_size)
 {
        /* set body_size to sequence contents size */
        body_size += 2 + (int) mech->len;         /* NEED overflow check */
-        return(1 + der_length_size(body_size) + body_size);
+        return 1 + der_length_size(body_size) + body_size;
 }
 EXPORT_SYMBOL_GPL(g_token_size);
@@ -186,27 +186,27 @@ g_verify_token_header(struct xdr_netobj *mech, int *body_size,
        int ret = 0;
        if ((toksize-=1) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if (*buf++ != 0x60)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if ((seqsize = der_read_length(&buf, &toksize)) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if (seqsize != toksize)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if ((toksize-=1) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if (*buf++ != 0x06)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if ((toksize-=1) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        toid.len = *buf++;
        if ((toksize-=toid.len) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        toid.data = buf;
        buf+=toid.len;
@@ -217,17 +217,17 @@ g_verify_token_header(struct xdr_netobj *mech, int *body_size,
      to return G_BAD_TOK_HEADER if the token header is in fact bad */
        if ((toksize-=2) < 0)
-                return(G_BAD_TOK_HEADER);
+                return G_BAD_TOK_HEADER;
        if (ret)
-                return(ret);
+                return ret;
        if (!ret) {
                *buf_in = buf;
                *body_size = toksize;
        }
-        return(ret);
+        return ret;
 }
 EXPORT_SYMBOL_GPL(g_verify_token_header);
diff --git a/net/sunrpc/auth_gss/gss_krb5_seqnum.c b/net/sunrpc/auth_gss/gss_krb5_seqnum.c
index 415c013ba382..62ac90c62cb1 100644
--- a/net/sunrpc/auth_gss/gss_krb5_seqnum.c
+++ b/net/sunrpc/auth_gss/gss_krb5_seqnum.c
@@ -162,5 +162,5 @@ krb5_get_seq_num(struct krb5_ctx *kctx,
        *seqnum = ((plain[0]) |
                   (plain[1] << 8) | (plain[2] << 16) | (plain[3] << 24));
-        return (0);
+        return 0;
 }
diff --git a/net/sunrpc/auth_gss/gss_mech_switch.c b/net/sunrpc/auth_gss/gss_mech_switch.c
index 2689de39dc78..8b4061049d76 100644
--- a/net/sunrpc/auth_gss/gss_mech_switch.c
+++ b/net/sunrpc/auth_gss/gss_mech_switch.c
@@ -331,7 +331,7 @@ gss_delete_sec_context(struct gss_ctx	**context_handle)
                        *context_handle);
        if (!*context_handle)
-                return(GSS_S_NO_CONTEXT);
+                return GSS_S_NO_CONTEXT;
        if ((*context_handle)->internal_ctx_id)
                (*context_handle)->mech_type->gm_ops
                        ->gss_delete_sec_context((*context_handle)
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index cace6049e4a5..aa5dbda6608c 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -376,7 +376,7 @@ int rpc_queue_empty(struct rpc_wait_queue *queue)
        spin_lock_bh(&queue->lock);
        res = queue->qlen;
        spin_unlock_bh(&queue->lock);
-        return (res == 0);
+        return res == 0;
 }
 EXPORT_SYMBOL_GPL(rpc_queue_empty);
diff --git a/net/tipc/addr.c b/net/tipc/addr.c
index c048543ffbeb..2ddc351b3be9 100644
--- a/net/tipc/addr.c
+++ b/net/tipc/addr.c
@@ -89,7 +89,7 @@ int tipc_addr_domain_valid(u32 addr)
 int tipc_addr_node_valid(u32 addr)
 {
-        return (tipc_addr_domain_valid(addr) && tipc_node(addr));
+        return tipc_addr_domain_valid(addr) && tipc_node(addr);
 }
 int tipc_in_scope(u32 domain, u32 addr)
diff --git a/net/tipc/bcast.c b/net/tipc/bcast.c
index a008c6689305..ecfaac10d0b4 100644
--- a/net/tipc/bcast.c
+++ b/net/tipc/bcast.c
@@ -143,6 +143,19 @@ static void bcbuf_decr_acks(struct sk_buff *buf)
 }
+static void bclink_set_last_sent(void)
+{
+        if (bcl->next_out)
+                bcl->fsm_msg_cnt = mod(buf_seqno(bcl->next_out) - 1);
+        else
+                bcl->fsm_msg_cnt = mod(bcl->next_out_no - 1);
+}
+u32 tipc_bclink_get_last_sent(void)
+{
+        return bcl->fsm_msg_cnt;
+}
 /**
 * bclink_set_gap - set gap according to contents of current deferred pkt queue
 *
@@ -171,7 +184,7 @@ static void bclink_set_gap(struct tipc_node *n_ptr)
 static int bclink_ack_allowed(u32 n)
 {
-        return((n % TIPC_MIN_LINK_WIN) == tipc_own_tag);
+        return (n % TIPC_MIN_LINK_WIN) == tipc_own_tag;
 }
@@ -237,8 +250,10 @@ void tipc_bclink_acknowledge(struct tipc_node *n_ptr, u32 acked)
        /* Try resolving broadcast link congestion, if necessary */
-        if (unlikely(bcl->next_out))
+        if (unlikely(bcl->next_out)) {
                tipc_link_push_queue(bcl);
+                bclink_set_last_sent();
+        }
        if (unlikely(released && !list_empty(&bcl->waiting_ports)))
                tipc_link_wakeup_ports(bcl, 0);
        spin_unlock_bh(&bc_lock);
@@ -395,7 +410,7 @@ int tipc_bclink_send_msg(struct sk_buff *buf)
        if (unlikely(res == -ELINKCONG))
                buf_discard(buf);
        else
-                bcl->stats.sent_info++;
+                bclink_set_last_sent();
        if (bcl->out_queue_size > bcl->stats.max_queue_sz)
                bcl->stats.max_queue_sz = bcl->out_queue_size;
@@ -529,15 +544,6 @@ receive:
        tipc_node_unlock(node);
 }
-u32 tipc_bclink_get_last_sent(void)
-{
-        u32 last_sent = mod(bcl->next_out_no - 1);
-        if (bcl->next_out)
-                last_sent = mod(buf_seqno(bcl->next_out) - 1);
-        return last_sent;
-}
 u32 tipc_bclink_acks_missing(struct tipc_node *n_ptr)
 {
        return (n_ptr->bclink.supported &&
@@ -570,6 +576,7 @@ static int tipc_bcbearer_send(struct sk_buff *buf,
                msg = buf_msg(buf);
                msg_set_non_seq(msg, 1);
                msg_set_mc_netid(msg, tipc_net_id);
+                bcl->stats.sent_info++;
        }
        /* Send buffer over bearers until all targets reached */
@@ -609,11 +616,13 @@ static int tipc_bcbearer_send(struct sk_buff *buf,
                bcbearer->remains = bcbearer->remains_new;
        }
-        /* Unable to reach all targets */
+        /*
+         * Unable to reach all targets (indicate success, since currently
+         * there isn't code in place to properly block & unblock the
+         * pseudo-bearer used by the broadcast link)
+         */
-        bcbearer->bearer.publ.blocked = 1;
+        return TIPC_OK;
-        bcl->stats.bearer_congs++;
-        return 1;
 }
 /**
diff --git a/net/tipc/bearer.c b/net/tipc/bearer.c
index 52ae17b2583e..9c10c6b7c12b 100644
--- a/net/tipc/bearer.c
+++ b/net/tipc/bearer.c
@@ -63,7 +63,7 @@ static int media_name_valid(const char *name)
        len = strlen(name);
        if ((len + 1) > TIPC_MAX_MEDIA_NAME)
                return 0;
-        return (strspn(name, tipc_alphabet) == len);
+        return strspn(name, tipc_alphabet) == len;
 }
 /**
diff --git a/net/tipc/core.c b/net/tipc/core.c
index 696468117985..466b861dab91 100644
--- a/net/tipc/core.c
+++ b/net/tipc/core.c
@@ -169,6 +169,7 @@ void tipc_core_stop(void)
        tipc_nametbl_stop();
        tipc_ref_table_stop();
        tipc_socket_stop();
+        tipc_log_resize(0);
 }
 /**
@@ -203,7 +204,9 @@ static int __init tipc_init(void)
 {
        int res;
-        tipc_log_resize(CONFIG_TIPC_LOG);
+        if (tipc_log_resize(CONFIG_TIPC_LOG) != 0)
+                warn("Unable to create log buffer\n");
        info("Activated (version " TIPC_MOD_VER
             " compiled " __DATE__ " " __TIME__ ")\n");
@@ -230,7 +233,6 @@ static void __exit tipc_exit(void)
        tipc_core_stop_net();
        tipc_core_stop();
        info("Deactivated\n");
-        tipc_log_resize(0);
 }
 module_init(tipc_init);
diff --git a/net/tipc/dbg.c b/net/tipc/dbg.c
index 1885a7edb0c8..6569d45bfb9a 100644
--- a/net/tipc/dbg.c
+++ b/net/tipc/dbg.c
@@ -134,7 +134,7 @@ void tipc_printbuf_reset(struct print_buf *pb)
 int tipc_printbuf_empty(struct print_buf *pb)
 {
-        return (!pb->buf || (pb->crs == pb->buf));
+        return !pb->buf || (pb->crs == pb->buf);
 }
 /**
@@ -169,7 +169,7 @@ int tipc_printbuf_validate(struct print_buf *pb)
                        tipc_printf(pb, err);
                }
        }
-        return (pb->crs - pb->buf + 1);
+        return pb->crs - pb->buf + 1;
 }
 /**
diff --git a/net/tipc/discover.c b/net/tipc/discover.c
index fc1fcf5e6b53..f28d1ae93125 100644
--- a/net/tipc/discover.c
+++ b/net/tipc/discover.c
@@ -203,6 +203,14 @@ void tipc_disc_recv_msg(struct sk_buff *buf, struct bearer *b_ptr)
                                return;
                }
                spin_lock_bh(&n_ptr->lock);
+                /* Don't talk to neighbor during cleanup after last session */
+                if (n_ptr->cleanup_required) {
+                        spin_unlock_bh(&n_ptr->lock);
+                        return;
+                }
                link = n_ptr->links[b_ptr->identity];
                if (!link) {
                        dbg("creating link\n");
diff --git a/net/tipc/eth_media.c b/net/tipc/eth_media.c
index 6230d16020c4..6e988ba485fd 100644
--- a/net/tipc/eth_media.c
+++ b/net/tipc/eth_media.c
@@ -72,17 +72,26 @@ static int send_msg(struct sk_buff *buf, struct tipc_bearer *tb_ptr,
 {
        struct sk_buff *clone;
        struct net_device *dev;
+        int delta;
        clone = skb_clone(buf, GFP_ATOMIC);
-        if (clone) {
+        if (!clone)
-                skb_reset_network_header(clone);
+                return 0;
-                dev = ((struct eth_bearer *)(tb_ptr->usr_handle))->dev;
-                clone->dev = dev;
+        dev = ((struct eth_bearer *)(tb_ptr->usr_handle))->dev;
-                dev_hard_header(clone, dev, ETH_P_TIPC,
+        delta = dev->hard_header_len - skb_headroom(buf);
-                                 &dest->dev_addr.eth_addr,
-                                 dev->dev_addr, clone->len);
+        if ((delta > 0) &&
-                dev_queue_xmit(clone);
+            pskb_expand_head(clone, SKB_DATA_ALIGN(delta), 0, GFP_ATOMIC)) {
+                kfree_skb(clone);
+                return 0;
        }
+        skb_reset_network_header(clone);
+        clone->dev = dev;
+        dev_hard_header(clone, dev, ETH_P_TIPC, &dest->dev_addr.eth_addr,
+                        dev->dev_addr, clone->len);
+        dev_queue_xmit(clone);
        return 0;
 }
@@ -92,15 +101,12 @@ static int send_msg(struct sk_buff *buf, struct tipc_bearer *tb_ptr,
 * Accept only packets explicitly sent to this node, or broadcast packets;
 * ignores packets sent using Ethernet multicast, and traffic sent to other
 * nodes (which can happen if interface is running in promiscuous mode).
- * Routine truncates any Ethernet padding/CRC appended to the message,
- * and ensures message size matches actual length
 */
 static int recv_msg(struct sk_buff *buf, struct net_device *dev,
                    struct packet_type *pt, struct net_device *orig_dev)
 {
        struct eth_bearer *eb_ptr = (struct eth_bearer *)pt->af_packet_priv;
-        u32 size;
        if (!net_eq(dev_net(dev), &init_net)) {
                kfree_skb(buf);
@@ -109,13 +115,9 @@ static int recv_msg(struct sk_buff *buf, struct net_device *dev,
        if (likely(eb_ptr->bearer)) {
                if (likely(buf->pkt_type <= PACKET_BROADCAST)) {
-                        size = msg_size((struct tipc_msg *)buf->data);
+                        buf->next = NULL;
-                        skb_trim(buf, size);
+                        tipc_recv_msg(buf, eb_ptr->bearer);
-                        if (likely(buf->len == size)) {
+                        return 0;
-                                buf->next = NULL;
-                                tipc_recv_msg(buf, eb_ptr->bearer);
-                                return 0;
-                        }
                }
        }
        kfree_skb(buf);
@@ -133,6 +135,16 @@ static int enable_bearer(struct tipc_bearer *tb_ptr)
        struct eth_bearer *eb_ptr = &eth_bearers[0];
        struct eth_bearer *stop = &eth_bearers[MAX_ETH_BEARERS];
        char *driver_name = strchr((const char *)tb_ptr->name, ':') + 1;
+        int pending_dev = 0;
+        /* Find unused Ethernet bearer structure */
+        while (eb_ptr->dev) {
+                if (!eb_ptr->bearer)
+                        pending_dev++;
+                if (++eb_ptr == stop)
+                        return pending_dev ? -EAGAIN : -EDQUOT;
+        }
        /* Find device with specified name */
diff --git a/net/tipc/link.c b/net/tipc/link.c
index a3616b99529b..b8cf1e9d0b86 100644
--- a/net/tipc/link.c
+++ b/net/tipc/link.c
@@ -239,13 +239,13 @@ int tipc_link_is_up(struct link *l_ptr)
 {
        if (!l_ptr)
                return 0;
-        return (link_working_working(l_ptr) || link_working_unknown(l_ptr));
+        return link_working_working(l_ptr) || link_working_unknown(l_ptr);
 }
 int tipc_link_is_active(struct link *l_ptr)
 {
-        return ((l_ptr->owner->active_links[0] == l_ptr) ||
+        return  (l_ptr->owner->active_links[0] == l_ptr) ||
-                (l_ptr->owner->active_links[1] == l_ptr));
+                (l_ptr->owner->active_links[1] == l_ptr);
 }
 /**
@@ -1802,6 +1802,15 @@ static int link_recv_buf_validate(struct sk_buff *buf)
        return pskb_may_pull(buf, hdr_size);
 }
+/**
+ * tipc_recv_msg - process TIPC messages arriving from off-node
+ * @head: pointer to message buffer chain
+ * @tb_ptr: pointer to bearer message arrived on
+ *
+ * Invoked with no locks held.  Bearer pointer must point to a valid bearer
+ * structure (i.e. cannot be NULL), but bearer can be inactive.
+ */
 void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr)
 {
        read_lock_bh(&tipc_net_lock);
@@ -1819,6 +1828,11 @@ void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr)
                head = head->next;
+                /* Ensure bearer is still enabled */
+                if (unlikely(!b_ptr->active))
+                        goto cont;
                /* Ensure message is well-formed */
                if (unlikely(!link_recv_buf_validate(buf)))
@@ -1855,13 +1869,22 @@ void tipc_recv_msg(struct sk_buff *head, struct tipc_bearer *tb_ptr)
                                goto cont;
                }
-                /* Locate unicast link endpoint that should handle message */
+                /* Locate neighboring node that sent message */
                n_ptr = tipc_node_find(msg_prevnode(msg));
                if (unlikely(!n_ptr))
                        goto cont;
                tipc_node_lock(n_ptr);
+                /* Don't talk to neighbor during cleanup after last session */
+                if (n_ptr->cleanup_required) {
+                        tipc_node_unlock(n_ptr);
+                        goto cont;
+                }
+                /* Locate unicast link endpoint that should handle message */
                l_ptr = n_ptr->links[b_ptr->identity];
                if (unlikely(!l_ptr)) {
                        tipc_node_unlock(n_ptr);
diff --git a/net/tipc/link.h b/net/tipc/link.h
index 2e5385c47d30..26151d30589d 100644
--- a/net/tipc/link.h
+++ b/net/tipc/link.h
@@ -279,12 +279,12 @@ static inline int between(u32 lower, u32 upper, u32 n)
 static inline int less_eq(u32 left, u32 right)
 {
-        return (mod(right - left) < 32768u);
+        return mod(right - left) < 32768u;
 }
 static inline int less(u32 left, u32 right)
 {
-        return (less_eq(left, right) && (mod(right) != mod(left)));
+        return less_eq(left, right) && (mod(right) != mod(left));
 }
 static inline u32 lesser(u32 left, u32 right)
@@ -299,32 +299,32 @@ static inline u32 lesser(u32 left, u32 right)
 static inline int link_working_working(struct link *l_ptr)
 {
-        return (l_ptr->state == WORKING_WORKING);
+        return l_ptr->state == WORKING_WORKING;
 }
 static inline int link_working_unknown(struct link *l_ptr)
 {
-        return (l_ptr->state == WORKING_UNKNOWN);
+        return l_ptr->state == WORKING_UNKNOWN;
 }
 static inline int link_reset_unknown(struct link *l_ptr)
 {
-        return (l_ptr->state == RESET_UNKNOWN);
+        return l_ptr->state == RESET_UNKNOWN;
 }
 static inline int link_reset_reset(struct link *l_ptr)
 {
-        return (l_ptr->state == RESET_RESET);
+        return l_ptr->state == RESET_RESET;
 }
 static inline int link_blocked(struct link *l_ptr)
 {
-        return (l_ptr->exp_msg_count || l_ptr->blocked);
+        return l_ptr->exp_msg_count || l_ptr->blocked;
 }
 static inline int link_congested(struct link *l_ptr)
 {
-        return (l_ptr->out_queue_size >= l_ptr->queue_limit[0]);
+        return l_ptr->out_queue_size >= l_ptr->queue_limit[0];
 }
 #endif
diff --git a/net/tipc/msg.h b/net/tipc/msg.h
index 995d2da35b01..031aad18efce 100644
--- a/net/tipc/msg.h
+++ b/net/tipc/msg.h
@@ -104,7 +104,7 @@ static inline u32 msg_user(struct tipc_msg *m)
 static inline u32 msg_isdata(struct tipc_msg *m)
 {
-        return (msg_user(m) <= TIPC_CRITICAL_IMPORTANCE);
+        return msg_user(m) <= TIPC_CRITICAL_IMPORTANCE;
 }
 static inline void msg_set_user(struct tipc_msg *m, u32 n)
@@ -289,7 +289,7 @@ static inline void msg_set_destnode(struct tipc_msg *m, u32 a)
 static inline int msg_is_dest(struct tipc_msg *m, u32 d)
 {
-        return(msg_short(m) || (msg_destnode(m) == d));
+        return msg_short(m) || (msg_destnode(m) == d);
 }
 static inline u32 msg_routed(struct tipc_msg *m)
@@ -632,7 +632,7 @@ static inline void msg_set_bcast_tag(struct tipc_msg *m, u32 n)
 static inline u32 msg_max_pkt(struct tipc_msg *m)
 {
-        return (msg_bits(m, 9, 16, 0xffff) * 4);
+        return msg_bits(m, 9, 16, 0xffff) * 4;
 }
 static inline void msg_set_max_pkt(struct tipc_msg *m, u32 n)
diff --git a/net/tipc/name_table.c b/net/tipc/name_table.c
index 8ba79620db3f..9ca4b0689237 100644
--- a/net/tipc/name_table.c
+++ b/net/tipc/name_table.c
@@ -116,7 +116,7 @@ DEFINE_RWLOCK(tipc_nametbl_lock);
 static int hash(int x)
 {
-        return(x & (tipc_nametbl_size - 1));
+        return x & (tipc_nametbl_size - 1);
 }
 /**
@@ -613,8 +613,7 @@ struct publication *tipc_nametbl_remove_publ(u32 type, u32 lower,
 }
 /*
- * tipc_nametbl_translate(): Translate tipc_name -> tipc_portid.
+ * tipc_nametbl_translate - translate name to port id
- *                      Very time-critical.
 *
 * Note: on entry 'destnode' is the search domain used during translation;
 *       on exit it passes back the node address of the matching port (if any)
@@ -685,7 +684,6 @@ found:
        }
        spin_unlock_bh(&seq->lock);
 not_found:
-        *destnode = 0;
        read_unlock_bh(&tipc_nametbl_lock);
        return 0;
 }
@@ -877,7 +875,7 @@ static void subseq_list(struct sub_seq *sseq, struct print_buf *buf, u32 depth,
                        u32 index)
 {
        char portIdStr[27];
-        char *scopeStr;
+        const char *scope_str[] = {"", " zone", " cluster", " node"};
        struct publication *publ = sseq->zone_list;
        tipc_printf(buf, "%-10u %-10u ", sseq->lower, sseq->upper);
@@ -893,15 +891,8 @@ static void subseq_list(struct sub_seq *sseq, struct print_buf *buf, u32 depth,
                         tipc_node(publ->node), publ->ref);
                tipc_printf(buf, "%-26s ", portIdStr);
                if (depth > 3) {
-                        if (publ->node != tipc_own_addr)
+                        tipc_printf(buf, "%-10u %s", publ->key,
-                                scopeStr = "";
+                                    scope_str[publ->scope]);
-                        else if (publ->scope == TIPC_NODE_SCOPE)
-                                scopeStr = "node";
-                        else if (publ->scope == TIPC_CLUSTER_SCOPE)
-                                scopeStr = "cluster";
-                        else
-                                scopeStr = "zone";
-                        tipc_printf(buf, "%-10u %s", publ->key, scopeStr);
                }
                publ = publ->zone_list_next;
@@ -951,24 +942,19 @@ static void nameseq_list(struct name_seq *seq, struct print_buf *buf, u32 depth,
 static void nametbl_header(struct print_buf *buf, u32 depth)
 {
-        tipc_printf(buf, "Type       ");
+        const char *header[] = {
+                "Type       ",
-        if (depth > 1)
+                "Lower      Upper      ",
-                tipc_printf(buf, "Lower      Upper      ");
+                "Port Identity              ",
-        if (depth > 2)
+                "Publication Scope"
-                tipc_printf(buf, "Port Identity              ");
+        };
-        if (depth > 3)
-                tipc_printf(buf, "Publication");
+        int i;
-        tipc_printf(buf, "\n-----------");
+        if (depth > 4)
+                depth = 4;
-        if (depth > 1)
+        for (i = 0; i < depth; i++)
-                tipc_printf(buf, "--------------------- ");
+                tipc_printf(buf, header[i]);
-        if (depth > 2)
-                tipc_printf(buf, "-------------------------- ");
-        if (depth > 3)
-                tipc_printf(buf, "------------------");
        tipc_printf(buf, "\n");
 }
diff --git a/net/tipc/net.c b/net/tipc/net.c
index f61b7694138b..7e05af47a196 100644
--- a/net/tipc/net.c
+++ b/net/tipc/net.c
@@ -248,6 +248,7 @@ void tipc_net_route_msg(struct sk_buff *buf)
        /* Handle message for another node */
        msg_dbg(msg, "NET>SEND>: ");
+        skb_trim(buf, msg_size(msg));
        tipc_link_send(buf, dnode, msg_link_selector(msg));
 }
diff --git a/net/tipc/node.c b/net/tipc/node.c
index b634942caba5..7c49cd056df7 100644
--- a/net/tipc/node.c
+++ b/net/tipc/node.c
@@ -237,23 +237,22 @@ void tipc_node_link_down(struct tipc_node *n_ptr, struct link *l_ptr)
 int tipc_node_has_active_links(struct tipc_node *n_ptr)
 {
-        return (n_ptr &&
+        return n_ptr->active_links[0] != NULL;
-                ((n_ptr->active_links[0]) || (n_ptr->active_links[1])));
 }
 int tipc_node_has_redundant_links(struct tipc_node *n_ptr)
 {
-        return (n_ptr->working_links > 1);
+        return n_ptr->working_links > 1;
 }
 static int tipc_node_has_active_routes(struct tipc_node *n_ptr)
 {
-        return (n_ptr && (n_ptr->last_router >= 0));
+        return n_ptr && (n_ptr->last_router >= 0);
 }
 int tipc_node_is_up(struct tipc_node *n_ptr)
 {
-        return (tipc_node_has_active_links(n_ptr) || tipc_node_has_active_routes(n_ptr));
+        return tipc_node_has_active_links(n_ptr) || tipc_node_has_active_routes(n_ptr);
 }
 struct tipc_node *tipc_node_attach_link(struct link *l_ptr)
@@ -384,6 +383,20 @@ static void node_established_contact(struct tipc_node *n_ptr)
                                  tipc_highest_allowed_slave);
 }
+static void node_cleanup_finished(unsigned long node_addr)
+{
+        struct tipc_node *n_ptr;
+        read_lock_bh(&tipc_net_lock);
+        n_ptr = tipc_node_find(node_addr);
+        if (n_ptr) {
+                tipc_node_lock(n_ptr);
+                n_ptr->cleanup_required = 0;
+                tipc_node_unlock(n_ptr);
+        }
+        read_unlock_bh(&tipc_net_lock);
+}
 static void node_lost_contact(struct tipc_node *n_ptr)
 {
        struct cluster *c_ptr;
@@ -458,6 +471,11 @@ static void node_lost_contact(struct tipc_node *n_ptr)
                tipc_k_signal((Handler)ns->handle_node_down,
                              (unsigned long)ns->usr_handle);
        }
+        /* Prevent re-contact with node until all cleanup is done */
+        n_ptr->cleanup_required = 1;
+        tipc_k_signal((Handler)node_cleanup_finished, n_ptr->addr);
 }
 /**
diff --git a/net/tipc/node.h b/net/tipc/node.h
index 6f990da5d143..45f3db3a595d 100644
--- a/net/tipc/node.h
+++ b/net/tipc/node.h
@@ -52,6 +52,7 @@
 * @active_links: pointers to active links to node
 * @links: pointers to all links to node
 * @working_links: number of working links to node (both active and standby)
+ * @cleanup_required: non-zero if cleaning up after a prior loss of contact
 * @link_cnt: number of links to node
 * @permit_changeover: non-zero if node has redundant links to this system
 * @routers: bitmap (used for multicluster communication)
@@ -78,6 +79,7 @@ struct tipc_node {
        struct link *links[MAX_BEARERS];
        int link_cnt;
        int working_links;
+        int cleanup_required;
        int permit_changeover;
        u32 routers[512/32];
        int last_router;
diff --git a/net/tipc/port.c b/net/tipc/port.c
index 0737680e9266..d760336f2ca8 100644
--- a/net/tipc/port.c
+++ b/net/tipc/port.c
@@ -588,19 +588,10 @@ void tipc_port_recv_proto_msg(struct sk_buff *buf)
        if (!p_ptr) {
                err = TIPC_ERR_NO_PORT;
        } else if (p_ptr->publ.connected) {
-                if (port_peernode(p_ptr) != msg_orignode(msg))
+                if ((port_peernode(p_ptr) != msg_orignode(msg)) ||
+                    (port_peerport(p_ptr) != msg_origport(msg))) {
                        err = TIPC_ERR_NO_PORT;
-                if (port_peerport(p_ptr) != msg_origport(msg))
+                } else if (msg_type(msg) == CONN_ACK) {
-                        err = TIPC_ERR_NO_PORT;
-                if (!err && msg_routed(msg)) {
-                        u32 seqno = msg_transp_seqno(msg);
-                        u32 myno =  ++p_ptr->last_in_seqno;
-                        if (seqno != myno) {
-                                err = TIPC_ERR_NO_PORT;
-                                abort_buf = port_build_self_abort_msg(p_ptr, err);
-                        }
-                }
-                if (msg_type(msg) == CONN_ACK) {
                        int wakeup = tipc_port_congested(p_ptr) &&
                                     p_ptr->publ.congested &&
                                     p_ptr->wakeup;
@@ -1473,7 +1464,7 @@ int tipc_forward2name(u32 ref,
        msg_set_destnode(msg, destnode);
        msg_set_destport(msg, destport);
-        if (likely(destport || destnode)) {
+        if (likely(destport)) {
                p_ptr->sent++;
                if (likely(destnode == tipc_own_addr))
                        return tipc_port_recv_sections(p_ptr, num_sect, msg_sect);
@@ -1551,7 +1542,7 @@ int tipc_forward_buf2name(u32 ref,
        skb_push(buf, LONG_H_SIZE);
        skb_copy_to_linear_data(buf, msg, LONG_H_SIZE);
        msg_dbg(buf_msg(buf),"PREP:");
-        if (likely(destport || destnode)) {
+        if (likely(destport)) {
                p_ptr->sent++;
                if (destnode == tipc_own_addr)
                        return tipc_port_recv_msg(buf);
diff --git a/net/tipc/port.h b/net/tipc/port.h
index 8d1652aab298..e74bd9563739 100644
--- a/net/tipc/port.h
+++ b/net/tipc/port.h
@@ -157,7 +157,7 @@ static inline u32 tipc_peer_node(struct port *p_ptr)
 static inline int tipc_port_congested(struct port *p_ptr)
 {
-        return((p_ptr->sent - p_ptr->acked) >= (TIPC_FLOW_CONTROL_WIN * 2));
+        return (p_ptr->sent - p_ptr->acked) >= (TIPC_FLOW_CONTROL_WIN * 2);
 }
 /**
diff --git a/net/tipc/socket.c b/net/tipc/socket.c
index 66e889ba48fd..33217fc3d697 100644
--- a/net/tipc/socket.c
+++ b/net/tipc/socket.c
@@ -64,6 +64,7 @@ struct tipc_sock {
        struct sock sk;
        struct tipc_port *p;
        struct tipc_portid peer_name;
+        long conn_timeout;
 };
 #define tipc_sk(sk) ((struct tipc_sock *)(sk))
@@ -240,9 +241,9 @@ static int tipc_create(struct net *net, struct socket *sock, int protocol,
        sock->state = state;
        sock_init_data(sock, sk);
-        sk->sk_rcvtimeo = msecs_to_jiffies(CONN_TIMEOUT_DEFAULT);
        sk->sk_backlog_rcv = backlog_rcv;
        tipc_sk(sk)->p = tp_ptr;
+        tipc_sk(sk)->conn_timeout = msecs_to_jiffies(CONN_TIMEOUT_DEFAULT);
        spin_unlock_bh(tp_ptr->lock);
@@ -429,36 +430,55 @@ static int get_name(struct socket *sock, struct sockaddr *uaddr,
 * to handle any preventable race conditions, so TIPC will do the same ...
 *
 * TIPC sets the returned events as follows:
- * a) POLLRDNORM and POLLIN are set if the socket's receive queue is non-empty
+ *
- *    or if a connection-oriented socket is does not have an active connection
+ * socket state         flags set
- *    (i.e. a read operation will not block).
+ * ------------         ---------
- * b) POLLOUT is set except when a socket's connection has been terminated
+ * unconnected          no read flags
- *    (i.e. a write operation will not block).
+ *                      no write flags
- * c) POLLHUP is set when a socket's connection has been terminated.
+ *
- *
+ * connecting           POLLIN/POLLRDNORM if ACK/NACK in rx queue
- * IMPORTANT: The fact that a read or write operation will not block does NOT
+ *                      no write flags
- * imply that the operation will succeed!
+ *
+ * connected            POLLIN/POLLRDNORM if data in rx queue
+ *                      POLLOUT if port is not congested
+ *
+ * disconnecting        POLLIN/POLLRDNORM/POLLHUP
+ *                      no write flags
+ *
+ * listening            POLLIN if SYN in rx queue
+ *                      no write flags
+ *
+ * ready                POLLIN/POLLRDNORM if data in rx queue
+ * [connectionless]     POLLOUT (since port cannot be congested)
+ *
+ * IMPORTANT: The fact that a read or write operation is indicated does NOT
+ * imply that the operation will succeed, merely that it should be performed
+ * and will not block.
 */
 static unsigned int poll(struct file *file, struct socket *sock,
                         poll_table *wait)
 {
        struct sock *sk = sock->sk;
-        u32 mask;
+        u32 mask = 0;
        poll_wait(file, sk_sleep(sk), wait);
-        if (!skb_queue_empty(&sk->sk_receive_queue) ||
+        switch ((int)sock->state) {
-            (sock->state == SS_UNCONNECTED) ||
+        case SS_READY:
-            (sock->state == SS_DISCONNECTING))
+        case SS_CONNECTED:
-                mask = (POLLRDNORM | POLLIN);
+                if (!tipc_sk_port(sk)->congested)
-        else
+                        mask |= POLLOUT;
-                mask = 0;
+                /* fall thru' */
+        case SS_CONNECTING:
-        if (sock->state == SS_DISCONNECTING)
+        case SS_LISTENING:
-                mask |= POLLHUP;
+                if (!skb_queue_empty(&sk->sk_receive_queue))
-        else
+                        mask |= (POLLIN | POLLRDNORM);
-                mask |= POLLOUT;
+                break;
+        case SS_DISCONNECTING:
+                mask = (POLLIN | POLLRDNORM | POLLHUP);
+                break;
+        }
        return mask;
 }
@@ -1026,9 +1046,8 @@ static int recv_stream(struct kiocb *iocb, struct socket *sock,
        struct sk_buff *buf;
        struct tipc_msg *msg;
        unsigned int sz;
-        int sz_to_copy;
+        int sz_to_copy, target, needed;
        int sz_copied = 0;
-        int needed;
        char __user *crs = m->msg_iov->iov_base;
        unsigned char *buf_crs;
        u32 err;
@@ -1050,6 +1069,8 @@ static int recv_stream(struct kiocb *iocb, struct socket *sock,
                goto exit;
        }
+        target = sock_rcvlowat(sk, flags & MSG_WAITALL, buf_len);
 restart:
        /* Look for a message in receive queue; wait if necessary */
@@ -1138,7 +1159,7 @@ restart:
        if ((sz_copied < buf_len) &&    /* didn't get all requested data */
            (!skb_queue_empty(&sk->sk_receive_queue) ||
-             (flags & MSG_WAITALL)) &&  /* and more is ready or required */
+            (sz_copied < target)) &&    /* and more is ready or required */
            (!(flags & MSG_PEEK)) &&    /* and aren't just peeking at data */
            (!err))                     /* and haven't reached a FIN */
                goto restart;
@@ -1174,7 +1195,7 @@ static int rx_queue_full(struct tipc_msg *msg, u32 queue_size, u32 base)
        if (msg_connected(msg))
                threshold *= 4;
-        return (queue_size >= threshold);
+        return queue_size >= threshold;
 }
 /**
@@ -1365,6 +1386,7 @@ static int connect(struct socket *sock, struct sockaddr *dest, int destlen,
        struct msghdr m = {NULL,};
        struct sk_buff *buf;
        struct tipc_msg *msg;
+        long timeout;
        int res;
        lock_sock(sk);
@@ -1379,7 +1401,7 @@ static int connect(struct socket *sock, struct sockaddr *dest, int destlen,
        /* For now, TIPC does not support the non-blocking form of connect() */
        if (flags & O_NONBLOCK) {
-                res = -EWOULDBLOCK;
+                res = -EOPNOTSUPP;
                goto exit;
        }
@@ -1425,11 +1447,12 @@ static int connect(struct socket *sock, struct sockaddr *dest, int destlen,
        /* Wait until an 'ACK' or 'RST' arrives, or a timeout occurs */
+        timeout = tipc_sk(sk)->conn_timeout;
        release_sock(sk);
        res = wait_event_interruptible_timeout(*sk_sleep(sk),
                        (!skb_queue_empty(&sk->sk_receive_queue) ||
                        (sock->state != SS_CONNECTING)),
-                        sk->sk_rcvtimeo);
+                        timeout ? timeout : MAX_SCHEDULE_TIMEOUT);
        lock_sock(sk);
        if (res > 0) {
@@ -1692,7 +1715,7 @@ static int setsockopt(struct socket *sock,
                res = tipc_set_portunreturnable(tport->ref, value);
                break;
        case TIPC_CONN_TIMEOUT:
-                sk->sk_rcvtimeo = msecs_to_jiffies(value);
+                tipc_sk(sk)->conn_timeout = msecs_to_jiffies(value);
                /* no need to set "res", since already 0 at this point */
                break;
        default:
@@ -1747,7 +1770,7 @@ static int getsockopt(struct socket *sock,
                res = tipc_portunreturnable(tport->ref, &value);
                break;
        case TIPC_CONN_TIMEOUT:
-                value = jiffies_to_msecs(sk->sk_rcvtimeo);
+                value = jiffies_to_msecs(tipc_sk(sk)->conn_timeout);
                /* no need to set "res", since already 0 at this point */
                break;
         case TIPC_NODE_RECVQ_DEPTH:
diff --git a/net/tipc/subscr.c b/net/tipc/subscr.c
index ab6eab4c45e2..1a5b9a6bd128 100644
--- a/net/tipc/subscr.c
+++ b/net/tipc/subscr.c
@@ -604,6 +604,6 @@ int tipc_ispublished(struct tipc_name const *name)
 {
        u32 domain = 0;
-        return(tipc_nametbl_translate(name->type, name->instance,&domain) != 0);
+        return tipc_nametbl_translate(name->type, name->instance, &domain) != 0;
 }
diff --git a/net/unix/af_unix.c b/net/unix/af_unix.c
index 0b39b2451ea5..c586da3f4f18 100644
--- a/net/unix/af_unix.c
+++ b/net/unix/af_unix.c
@@ -2033,11 +2033,10 @@ static unsigned int unix_poll(struct file *file, struct socket *sock, poll_table
        if (sk->sk_shutdown == SHUTDOWN_MASK)
                mask |= POLLHUP;
        if (sk->sk_shutdown & RCV_SHUTDOWN)
-                mask |= POLLRDHUP;
+                mask |= POLLRDHUP | POLLIN | POLLRDNORM;
        /* readable? */
-        if (!skb_queue_empty(&sk->sk_receive_queue) ||
+        if (!skb_queue_empty(&sk->sk_receive_queue))
-            (sk->sk_shutdown & RCV_SHUTDOWN))
                mask |= POLLIN | POLLRDNORM;
        /* Connection-based need to check for termination and startup */
diff --git a/net/wireless/core.c b/net/wireless/core.c
index d6d046b9f6f2..9c21ebf9780e 100644
--- a/net/wireless/core.c
+++ b/net/wireless/core.c
@@ -253,11 +253,16 @@ int cfg80211_switch_netns(struct cfg80211_registered_device *rdev,
                        WARN_ON(err);
                        wdev->netdev->features |= NETIF_F_NETNS_LOCAL;
                }
+                return err;
        }
        wiphy_net_set(&rdev->wiphy, net);
-        return err;
+        err = device_rename(&rdev->wiphy.dev, dev_name(&rdev->wiphy.dev));
+        WARN_ON(err);
+        return 0;
 }
 static void cfg80211_rfkill_poll(struct rfkill *rfkill, void *data)
@@ -428,7 +433,7 @@ int wiphy_register(struct wiphy *wiphy)
        /* sanity check ifmodes */
        WARN_ON(!ifmodes);
-        ifmodes &= ((1 << __NL80211_IFTYPE_AFTER_LAST) - 1) & ~1;
+        ifmodes &= ((1 << NUM_NL80211_IFTYPES) - 1) & ~1;
        if (WARN_ON(ifmodes != wiphy->interface_modes))
                wiphy->interface_modes = ifmodes;
@@ -683,8 +688,8 @@ static int cfg80211_netdev_notifier_call(struct notifier_block * nb,
                INIT_WORK(&wdev->cleanup_work, wdev_cleanup_work);
                INIT_LIST_HEAD(&wdev->event_list);
                spin_lock_init(&wdev->event_lock);
-                INIT_LIST_HEAD(&wdev->action_registrations);
+                INIT_LIST_HEAD(&wdev->mgmt_registrations);
-                spin_lock_init(&wdev->action_registrations_lock);
+                spin_lock_init(&wdev->mgmt_registrations_lock);
                mutex_lock(&rdev->devlist_mtx);
                list_add_rcu(&wdev->list, &rdev->netdev_list);
@@ -724,6 +729,7 @@ static int cfg80211_netdev_notifier_call(struct notifier_block * nb,
                        dev->ethtool_ops = &cfg80211_ethtool_ops;
                if ((wdev->iftype == NL80211_IFTYPE_STATION ||
+                     wdev->iftype == NL80211_IFTYPE_P2P_CLIENT ||
                     wdev->iftype == NL80211_IFTYPE_ADHOC) && !wdev->use_4addr)
                        dev->priv_flags |= IFF_DONT_BRIDGE;
                break;
@@ -732,6 +738,7 @@ static int cfg80211_netdev_notifier_call(struct notifier_block * nb,
                case NL80211_IFTYPE_ADHOC:
                        cfg80211_leave_ibss(rdev, dev, true);
                        break;
+                case NL80211_IFTYPE_P2P_CLIENT:
                case NL80211_IFTYPE_STATION:
                        wdev_lock(wdev);
 #ifdef CONFIG_CFG80211_WEXT
@@ -804,7 +811,7 @@ static int cfg80211_netdev_notifier_call(struct notifier_block * nb,
                        sysfs_remove_link(&dev->dev.kobj, "phy80211");
                        list_del_rcu(&wdev->list);
                        rdev->devlist_generation++;
-                        cfg80211_mlme_purge_actions(wdev);
+                        cfg80211_mlme_purge_registrations(wdev);
 #ifdef CONFIG_CFG80211_WEXT
                        kfree(wdev->wext.keys);
 #endif
@@ -910,52 +917,3 @@ static void __exit cfg80211_exit(void)
        destroy_workqueue(cfg80211_wq);
 }
 module_exit(cfg80211_exit);
-static int ___wiphy_printk(const char *level, const struct wiphy *wiphy,
-                           struct va_format *vaf)
-{
-        if (!wiphy)
-                return printk("%s(NULL wiphy *): %pV", level, vaf);
-        return printk("%s%s: %pV", level, wiphy_name(wiphy), vaf);
-}
-int __wiphy_printk(const char *level, const struct wiphy *wiphy,
-                   const char *fmt, ...)
-{
-        struct va_format vaf;
-        va_list args;
-        int r;
-        va_start(args, fmt);
-        vaf.fmt = fmt;
-        vaf.va = &args;
-        r = ___wiphy_printk(level, wiphy, &vaf);
-        va_end(args);
-        return r;
-}
-EXPORT_SYMBOL(__wiphy_printk);
-#define define_wiphy_printk_level(func, kern_level)             \
-int func(const struct wiphy *wiphy, const char *fmt, ...)       \
-{                                                               \
-        struct va_format vaf;                                   \
-        va_list args;                                           \
-        int r;                                                  \
-                                                                \
-        va_start(args, fmt);                                    \
-                                                                \
-        vaf.fmt = fmt;                                          \
-        vaf.va = &args;                                         \
-                                                                \
-        r = ___wiphy_printk(kern_level, wiphy, &vaf);           \
-        va_end(args);                                           \
-                                                                \
-        return r;                                               \
-}                                                               \
-EXPORT_SYMBOL(func);
-define_wiphy_printk_level(wiphy_debug, KERN_DEBUG);
diff --git a/net/wireless/core.h b/net/wireless/core.h
index 63d57ae399c3..5d89310b3587 100644
--- a/net/wireless/core.h
+++ b/net/wireless/core.h
@@ -86,7 +86,7 @@ struct cfg80211_registered_device *wiphy_to_dev(struct wiphy *wiphy)
 static inline
 bool wiphy_idx_valid(int wiphy_idx)
 {
-        return (wiphy_idx >= 0);
+        return wiphy_idx >= 0;
 }
@@ -95,7 +95,10 @@ extern struct mutex cfg80211_mutex;
 extern struct list_head cfg80211_rdev_list;
 extern int cfg80211_rdev_list_generation;
-#define assert_cfg80211_lock() WARN_ON(!mutex_is_locked(&cfg80211_mutex))
+static inline void assert_cfg80211_lock(void)
+{
+        lockdep_assert_held(&cfg80211_mutex);
+}
 /*
 * You can use this to mark a wiphy_idx as not having an associated wiphy.
@@ -202,8 +205,8 @@ static inline void wdev_unlock(struct wireless_dev *wdev)
        mutex_unlock(&wdev->mtx);
 }
-#define ASSERT_RDEV_LOCK(rdev) WARN_ON(!mutex_is_locked(&(rdev)->mtx));
+#define ASSERT_RDEV_LOCK(rdev) lockdep_assert_held(&(rdev)->mtx)
-#define ASSERT_WDEV_LOCK(wdev) WARN_ON(!mutex_is_locked(&(wdev)->mtx));
+#define ASSERT_WDEV_LOCK(wdev) lockdep_assert_held(&(wdev)->mtx)
 enum cfg80211_event_type {
        EVENT_CONNECT_RESULT,
@@ -331,16 +334,17 @@ void __cfg80211_connect_result(struct net_device *dev, const u8 *bssid,
                               const u8 *resp_ie, size_t resp_ie_len,
                               u16 status, bool wextev,
                               struct cfg80211_bss *bss);
-int cfg80211_mlme_register_action(struct wireless_dev *wdev, u32 snd_pid,
+int cfg80211_mlme_register_mgmt(struct wireless_dev *wdev, u32 snd_pid,
-                                  const u8 *match_data, int match_len);
+                                u16 frame_type, const u8 *match_data,
-void cfg80211_mlme_unregister_actions(struct wireless_dev *wdev, u32 nlpid);
+                                int match_len);
-void cfg80211_mlme_purge_actions(struct wireless_dev *wdev);
+void cfg80211_mlme_unregister_socket(struct wireless_dev *wdev, u32 nlpid);
-int cfg80211_mlme_action(struct cfg80211_registered_device *rdev,
+void cfg80211_mlme_purge_registrations(struct wireless_dev *wdev);
-                         struct net_device *dev,
+int cfg80211_mlme_mgmt_tx(struct cfg80211_registered_device *rdev,
-                         struct ieee80211_channel *chan,
+                          struct net_device *dev,
-                         enum nl80211_channel_type channel_type,
+                          struct ieee80211_channel *chan,
-                         bool channel_type_valid,
+                          enum nl80211_channel_type channel_type,
-                         const u8 *buf, size_t len, u64 *cookie);
+                          bool channel_type_valid,
+                          const u8 *buf, size_t len, u64 *cookie);
 /* SME */
 int __cfg80211_connect(struct cfg80211_registered_device *rdev,
diff --git a/net/wireless/ibss.c b/net/wireless/ibss.c
index 27a8ce9343c3..8cb6e08373b9 100644
--- a/net/wireless/ibss.c
+++ b/net/wireless/ibss.c
@@ -88,6 +88,25 @@ int __cfg80211_join_ibss(struct cfg80211_registered_device *rdev,
        if (wdev->ssid_len)
                return -EALREADY;
+        if (!params->basic_rates) {
+                /*
+                * If no rates were explicitly configured,
+                * use the mandatory rate set for 11b or
+                * 11a for maximum compatibility.
+                */
+                struct ieee80211_supported_band *sband =
+                        rdev->wiphy.bands[params->channel->band];
+                int j;
+                u32 flag = params->channel->band == IEEE80211_BAND_5GHZ ?
+                        IEEE80211_RATE_MANDATORY_A :
+                        IEEE80211_RATE_MANDATORY_B;
+                for (j = 0; j < sband->n_bitrates; j++) {
+                        if (sband->bitrates[j].flags & flag)
+                                params->basic_rates |= BIT(j);
+                }
+        }
        if (WARN_ON(wdev->connect_keys))
                kfree(wdev->connect_keys);
        wdev->connect_keys = connkeys;
diff --git a/net/wireless/mlme.c b/net/wireless/mlme.c
index d1a3fb99fdf2..46f371160896 100644
--- a/net/wireless/mlme.c
+++ b/net/wireless/mlme.c
@@ -149,7 +149,7 @@ void __cfg80211_send_deauth(struct net_device *dev,
        struct ieee80211_mgmt *mgmt = (struct ieee80211_mgmt *)buf;
        const u8 *bssid = mgmt->bssid;
        int i;
-        bool found = false;
+        bool found = false, was_current = false;
        ASSERT_WDEV_LOCK(wdev);
@@ -159,6 +159,7 @@ void __cfg80211_send_deauth(struct net_device *dev,
                cfg80211_put_bss(&wdev->current_bss->pub);
                wdev->current_bss = NULL;
                found = true;
+                was_current = true;
        } else for (i = 0; i < MAX_AUTH_BSSES; i++) {
                if (wdev->auth_bsses[i] &&
                    memcmp(wdev->auth_bsses[i]->pub.bssid, bssid, ETH_ALEN) == 0) {
@@ -183,7 +184,7 @@ void __cfg80211_send_deauth(struct net_device *dev,
        nl80211_send_deauth(rdev, dev, buf, len, GFP_KERNEL);
-        if (wdev->sme_state == CFG80211_SME_CONNECTED) {
+        if (wdev->sme_state == CFG80211_SME_CONNECTED && was_current) {
                u16 reason_code;
                bool from_ap;
@@ -747,31 +748,51 @@ void cfg80211_new_sta(struct net_device *dev, const u8 *mac_addr,
 }
 EXPORT_SYMBOL(cfg80211_new_sta);
-struct cfg80211_action_registration {
+struct cfg80211_mgmt_registration {
        struct list_head list;
        u32 nlpid;
        int match_len;
+        __le16 frame_type;
        u8 match[];
 };
-int cfg80211_mlme_register_action(struct wireless_dev *wdev, u32 snd_pid,
+int cfg80211_mlme_register_mgmt(struct wireless_dev *wdev, u32 snd_pid,
-                                  const u8 *match_data, int match_len)
+                                u16 frame_type, const u8 *match_data,
+                                int match_len)
 {
-        struct cfg80211_action_registration *reg, *nreg;
+        struct cfg80211_mgmt_registration *reg, *nreg;
        int err = 0;
+        u16 mgmt_type;
+        if (!wdev->wiphy->mgmt_stypes)
+                return -EOPNOTSUPP;
+        if ((frame_type & IEEE80211_FCTL_FTYPE) != IEEE80211_FTYPE_MGMT)
+                return -EINVAL;
+        if (frame_type & ~(IEEE80211_FCTL_FTYPE | IEEE80211_FCTL_STYPE))
+                return -EINVAL;
+        mgmt_type = (frame_type & IEEE80211_FCTL_STYPE) >> 4;
+        if (!(wdev->wiphy->mgmt_stypes[wdev->iftype].rx & BIT(mgmt_type)))
+                return -EINVAL;
        nreg = kzalloc(sizeof(*reg) + match_len, GFP_KERNEL);
        if (!nreg)
                return -ENOMEM;
-        spin_lock_bh(&wdev->action_registrations_lock);
+        spin_lock_bh(&wdev->mgmt_registrations_lock);
-        list_for_each_entry(reg, &wdev->action_registrations, list) {
+        list_for_each_entry(reg, &wdev->mgmt_registrations, list) {
                int mlen = min(match_len, reg->match_len);
+                if (frame_type != le16_to_cpu(reg->frame_type))
+                        continue;
                if (memcmp(reg->match, match_data, mlen) == 0) {
                        err = -EALREADY;
                        break;
@@ -786,69 +807,83 @@ int cfg80211_mlme_register_action(struct wireless_dev *wdev, u32 snd_pid,
        memcpy(nreg->match, match_data, match_len);
        nreg->match_len = match_len;
        nreg->nlpid = snd_pid;
-        list_add(&nreg->list, &wdev->action_registrations);
+        nreg->frame_type = cpu_to_le16(frame_type);
+        list_add(&nreg->list, &wdev->mgmt_registrations);
 out:
-        spin_unlock_bh(&wdev->action_registrations_lock);
+        spin_unlock_bh(&wdev->mgmt_registrations_lock);
        return err;
 }
-void cfg80211_mlme_unregister_actions(struct wireless_dev *wdev, u32 nlpid)
+void cfg80211_mlme_unregister_socket(struct wireless_dev *wdev, u32 nlpid)
 {
-        struct cfg80211_action_registration *reg, *tmp;
+        struct cfg80211_mgmt_registration *reg, *tmp;
-        spin_lock_bh(&wdev->action_registrations_lock);
+        spin_lock_bh(&wdev->mgmt_registrations_lock);
-        list_for_each_entry_safe(reg, tmp, &wdev->action_registrations, list) {
+        list_for_each_entry_safe(reg, tmp, &wdev->mgmt_registrations, list) {
                if (reg->nlpid == nlpid) {
                        list_del(&reg->list);
                        kfree(reg);
                }
        }
-        spin_unlock_bh(&wdev->action_registrations_lock);
+        spin_unlock_bh(&wdev->mgmt_registrations_lock);
 }
-void cfg80211_mlme_purge_actions(struct wireless_dev *wdev)
+void cfg80211_mlme_purge_registrations(struct wireless_dev *wdev)
 {
-        struct cfg80211_action_registration *reg, *tmp;
+        struct cfg80211_mgmt_registration *reg, *tmp;
-        spin_lock_bh(&wdev->action_registrations_lock);
+        spin_lock_bh(&wdev->mgmt_registrations_lock);
-        list_for_each_entry_safe(reg, tmp, &wdev->action_registrations, list) {
+        list_for_each_entry_safe(reg, tmp, &wdev->mgmt_registrations, list) {
                list_del(&reg->list);
                kfree(reg);
        }
-        spin_unlock_bh(&wdev->action_registrations_lock);
+        spin_unlock_bh(&wdev->mgmt_registrations_lock);
 }
-int cfg80211_mlme_action(struct cfg80211_registered_device *rdev,
+int cfg80211_mlme_mgmt_tx(struct cfg80211_registered_device *rdev,
-                         struct net_device *dev,
+                          struct net_device *dev,
-                         struct ieee80211_channel *chan,
+                          struct ieee80211_channel *chan,
-                         enum nl80211_channel_type channel_type,
+                          enum nl80211_channel_type channel_type,
-                         bool channel_type_valid,
+                          bool channel_type_valid,
-                         const u8 *buf, size_t len, u64 *cookie)
+                          const u8 *buf, size_t len, u64 *cookie)
 {
        struct wireless_dev *wdev = dev->ieee80211_ptr;
        const struct ieee80211_mgmt *mgmt;
+        u16 stype;
+        if (!wdev->wiphy->mgmt_stypes)
+                return -EOPNOTSUPP;
-        if (rdev->ops->action == NULL)
+        if (!rdev->ops->mgmt_tx)
                return -EOPNOTSUPP;
        if (len < 24 + 1)
                return -EINVAL;
        mgmt = (const struct ieee80211_mgmt *) buf;
-        if (!ieee80211_is_action(mgmt->frame_control))
+        if (!ieee80211_is_mgmt(mgmt->frame_control))
                return -EINVAL;
-        if (mgmt->u.action.category != WLAN_CATEGORY_PUBLIC) {
+        stype = le16_to_cpu(mgmt->frame_control) & IEEE80211_FCTL_STYPE;
+        if (!(wdev->wiphy->mgmt_stypes[wdev->iftype].tx & BIT(stype >> 4)))
+                return -EINVAL;
+        if (ieee80211_is_action(mgmt->frame_control) &&
+            mgmt->u.action.category != WLAN_CATEGORY_PUBLIC) {
                /* Verify that we are associated with the destination AP */
                wdev_lock(wdev);
                if (!wdev->current_bss ||
                    memcmp(wdev->current_bss->pub.bssid, mgmt->bssid,
                           ETH_ALEN) != 0 ||
-                    (wdev->iftype == NL80211_IFTYPE_STATION &&
+                    ((wdev->iftype == NL80211_IFTYPE_STATION ||
+                      wdev->iftype == NL80211_IFTYPE_P2P_CLIENT) &&
                     memcmp(wdev->current_bss->pub.bssid, mgmt->da,
                            ETH_ALEN) != 0)) {
                        wdev_unlock(wdev);
@@ -862,64 +897,75 @@ int cfg80211_mlme_action(struct cfg80211_registered_device *rdev,
                return -EINVAL;
        /* Transmit the Action frame as requested by user space */
-        return rdev->ops->action(&rdev->wiphy, dev, chan, channel_type,
+        return rdev->ops->mgmt_tx(&rdev->wiphy, dev, chan, channel_type,
-                                 channel_type_valid, buf, len, cookie);
+                                  channel_type_valid, buf, len, cookie);
 }
-bool cfg80211_rx_action(struct net_device *dev, int freq, const u8 *buf,
+bool cfg80211_rx_mgmt(struct net_device *dev, int freq, const u8 *buf,
-                        size_t len, gfp_t gfp)
+                      size_t len, gfp_t gfp)
 {
        struct wireless_dev *wdev = dev->ieee80211_ptr;
        struct wiphy *wiphy = wdev->wiphy;
        struct cfg80211_registered_device *rdev = wiphy_to_dev(wiphy);
-        struct cfg80211_action_registration *reg;
+        struct cfg80211_mgmt_registration *reg;
-        const u8 *action_data;
+        const struct ieee80211_txrx_stypes *stypes =
-        int action_data_len;
+                &wiphy->mgmt_stypes[wdev->iftype];
+        struct ieee80211_mgmt *mgmt = (void *)buf;
+        const u8 *data;
+        int data_len;
        bool result = false;
+        __le16 ftype = mgmt->frame_control &
+                cpu_to_le16(IEEE80211_FCTL_FTYPE | IEEE80211_FCTL_STYPE);
+        u16 stype;
-        /* frame length - min size excluding category */
+        stype = (le16_to_cpu(mgmt->frame_control) & IEEE80211_FCTL_STYPE) >> 4;
-        action_data_len = len - (IEEE80211_MIN_ACTION_SIZE - 1);
-        /* action data starts with category */
+        if (!(stypes->rx & BIT(stype)))
-        action_data = buf + IEEE80211_MIN_ACTION_SIZE - 1;
+                return false;
-        spin_lock_bh(&wdev->action_registrations_lock);
+        data = buf + ieee80211_hdrlen(mgmt->frame_control);
+        data_len = len - ieee80211_hdrlen(mgmt->frame_control);
+        spin_lock_bh(&wdev->mgmt_registrations_lock);
+        list_for_each_entry(reg, &wdev->mgmt_registrations, list) {
+                if (reg->frame_type != ftype)
+                        continue;
-        list_for_each_entry(reg, &wdev->action_registrations, list) {
+                if (reg->match_len > data_len)
-                if (reg->match_len > action_data_len)
                        continue;
-                if (memcmp(reg->match, action_data, reg->match_len))
+                if (memcmp(reg->match, data, reg->match_len))
                        continue;
                /* found match! */
                /* Indicate the received Action frame to user space */
-                if (nl80211_send_action(rdev, dev, reg->nlpid, freq,
+                if (nl80211_send_mgmt(rdev, dev, reg->nlpid, freq,
-                                        buf, len, gfp))
+                                      buf, len, gfp))
                        continue;
                result = true;
                break;
        }
-        spin_unlock_bh(&wdev->action_registrations_lock);
+        spin_unlock_bh(&wdev->mgmt_registrations_lock);
        return result;
 }
-EXPORT_SYMBOL(cfg80211_rx_action);
+EXPORT_SYMBOL(cfg80211_rx_mgmt);
-void cfg80211_action_tx_status(struct net_device *dev, u64 cookie,
+void cfg80211_mgmt_tx_status(struct net_device *dev, u64 cookie,
-                               const u8 *buf, size_t len, bool ack, gfp_t gfp)
+                             const u8 *buf, size_t len, bool ack, gfp_t gfp)
 {
        struct wireless_dev *wdev = dev->ieee80211_ptr;
        struct wiphy *wiphy = wdev->wiphy;
        struct cfg80211_registered_device *rdev = wiphy_to_dev(wiphy);
        /* Indicate TX status of the Action frame to user space */
-        nl80211_send_action_tx_status(rdev, dev, cookie, buf, len, ack, gfp);
+        nl80211_send_mgmt_tx_status(rdev, dev, cookie, buf, len, ack, gfp);
 }
-EXPORT_SYMBOL(cfg80211_action_tx_status);
+EXPORT_SYMBOL(cfg80211_mgmt_tx_status);
 void cfg80211_cqm_rssi_notify(struct net_device *dev,
                              enum nl80211_cqm_rssi_threshold_event rssi_event,
diff --git a/net/wireless/nl80211.c b/net/wireless/nl80211.c
index 37902a54e9c1..9c84825803ce 100644
--- a/net/wireless/nl80211.c
+++ b/net/wireless/nl80211.c
@@ -136,6 +136,8 @@ static const struct nla_policy nl80211_policy[NL80211_ATTR_MAX+1] = {
                .len = sizeof(struct nl80211_sta_flag_update),
        },
        [NL80211_ATTR_CONTROL_PORT] = { .type = NLA_FLAG },
+        [NL80211_ATTR_CONTROL_PORT_ETHERTYPE] = { .type = NLA_U16 },
+        [NL80211_ATTR_CONTROL_PORT_NO_ENCRYPT] = { .type = NLA_FLAG },
        [NL80211_ATTR_PRIVACY] = { .type = NLA_FLAG },
        [NL80211_ATTR_CIPHER_SUITE_GROUP] = { .type = NLA_U32 },
        [NL80211_ATTR_WPA_VERSIONS] = { .type = NLA_U32 },
@@ -156,6 +158,7 @@ static const struct nla_policy nl80211_policy[NL80211_ATTR_MAX+1] = {
        [NL80211_ATTR_WIPHY_TX_POWER_SETTING] = { .type = NLA_U32 },
        [NL80211_ATTR_WIPHY_TX_POWER_LEVEL] = { .type = NLA_U32 },
+        [NL80211_ATTR_FRAME_TYPE] = { .type = NLA_U16 },
 };
 /* policy for the attributes */
@@ -407,12 +410,14 @@ static int nl80211_key_allowed(struct wireless_dev *wdev)
        switch (wdev->iftype) {
        case NL80211_IFTYPE_AP:
        case NL80211_IFTYPE_AP_VLAN:
+        case NL80211_IFTYPE_P2P_GO:
                break;
        case NL80211_IFTYPE_ADHOC:
                if (!wdev->current_bss)
                        return -ENOLINK;
                break;
        case NL80211_IFTYPE_STATION:
+        case NL80211_IFTYPE_P2P_CLIENT:
                if (wdev->sme_state != CFG80211_SME_CONNECTED)
                        return -ENOLINK;
                break;
@@ -437,6 +442,8 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 pid, u32 seq, int flags,
        struct ieee80211_rate *rate;
        int i;
        u16 ifmodes = dev->wiphy.interface_modes;
+        const struct ieee80211_txrx_stypes *mgmt_stypes =
+                                dev->wiphy.mgmt_stypes;
        hdr = nl80211hdr_put(msg, pid, seq, flags, NL80211_CMD_NEW_WIPHY);
        if (!hdr)
@@ -471,6 +478,9 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 pid, u32 seq, int flags,
        NLA_PUT_U8(msg, NL80211_ATTR_MAX_NUM_PMKIDS,
                   dev->wiphy.max_num_pmkids);
+        if (dev->wiphy.flags & WIPHY_FLAG_CONTROL_PORT_PROTOCOL)
+                NLA_PUT_FLAG(msg, NL80211_ATTR_CONTROL_PORT_ETHERTYPE);
        nl_modes = nla_nest_start(msg, NL80211_ATTR_SUPPORTED_IFTYPES);
        if (!nl_modes)
                goto nla_put_failure;
@@ -587,7 +597,7 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 pid, u32 seq, int flags,
        CMD(flush_pmksa, FLUSH_PMKSA);
        CMD(remain_on_channel, REMAIN_ON_CHANNEL);
        CMD(set_bitrate_mask, SET_TX_BITRATE_MASK);
-        CMD(action, ACTION);
+        CMD(mgmt_tx, FRAME);
        if (dev->wiphy.flags & WIPHY_FLAG_NETNS_OK) {
                i++;
                NLA_PUT_U32(msg, i, NL80211_CMD_SET_WIPHY_NETNS);
@@ -608,6 +618,55 @@ static int nl80211_send_wiphy(struct sk_buff *msg, u32 pid, u32 seq, int flags,
        nla_nest_end(msg, nl_cmds);
+        if (mgmt_stypes) {
+                u16 stypes;
+                struct nlattr *nl_ftypes, *nl_ifs;
+                enum nl80211_iftype ift;
+                nl_ifs = nla_nest_start(msg, NL80211_ATTR_TX_FRAME_TYPES);
+                if (!nl_ifs)
+                        goto nla_put_failure;
+                for (ift = 0; ift < NUM_NL80211_IFTYPES; ift++) {
+                        nl_ftypes = nla_nest_start(msg, ift);
+                        if (!nl_ftypes)
+                                goto nla_put_failure;
+                        i = 0;
+                        stypes = mgmt_stypes[ift].tx;
+                        while (stypes) {
+                                if (stypes & 1)
+                                        NLA_PUT_U16(msg, NL80211_ATTR_FRAME_TYPE,
+                                                    (i << 4) | IEEE80211_FTYPE_MGMT);
+                                stypes >>= 1;
+                                i++;
+                        }
+                        nla_nest_end(msg, nl_ftypes);
+                }
+                nla_nest_end(msg, nl_ifs);
+                nl_ifs = nla_nest_start(msg, NL80211_ATTR_RX_FRAME_TYPES);
+                if (!nl_ifs)
+                        goto nla_put_failure;
+                for (ift = 0; ift < NUM_NL80211_IFTYPES; ift++) {
+                        nl_ftypes = nla_nest_start(msg, ift);
+                        if (!nl_ftypes)
+                                goto nla_put_failure;
+                        i = 0;
+                        stypes = mgmt_stypes[ift].rx;
+                        while (stypes) {
+                                if (stypes & 1)
+                                        NLA_PUT_U16(msg, NL80211_ATTR_FRAME_TYPE,
+                                                    (i << 4) | IEEE80211_FTYPE_MGMT);
+                                stypes >>= 1;
+                                i++;
+                        }
+                        nla_nest_end(msg, nl_ftypes);
+                }
+                nla_nest_end(msg, nl_ifs);
+        }
        return genlmsg_end(msg, hdr);
 nla_put_failure:
@@ -709,7 +768,8 @@ static bool nl80211_can_set_dev_channel(struct wireless_dev *wdev)
                wdev->iftype == NL80211_IFTYPE_AP ||
                wdev->iftype == NL80211_IFTYPE_WDS ||
                wdev->iftype == NL80211_IFTYPE_MESH_POINT ||
-                wdev->iftype == NL80211_IFTYPE_MONITOR;
+                wdev->iftype == NL80211_IFTYPE_MONITOR ||
+                wdev->iftype == NL80211_IFTYPE_P2P_GO;
 }
 static int __nl80211_set_channel(struct cfg80211_registered_device *rdev,
@@ -776,7 +836,7 @@ static int nl80211_set_wiphy(struct sk_buff *skb, struct genl_info *info)
        struct cfg80211_registered_device *rdev;
        struct net_device *netdev = NULL;
        struct wireless_dev *wdev;
-        int result, rem_txq_params = 0;
+        int result = 0, rem_txq_params = 0;
        struct nlattr *nl_txq_params;
        u32 changed;
        u8 retry_short = 0, retry_long = 0;
@@ -1636,7 +1696,8 @@ static int nl80211_addset_beacon(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto unlock_rtnl;
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_GO) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -1728,7 +1789,8 @@ static int nl80211_del_beacon(struct sk_buff *skb, struct genl_info *info)
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_GO) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -2071,10 +2133,12 @@ static int nl80211_set_station(struct sk_buff *skb, struct genl_info *info)
        switch (dev->ieee80211_ptr->iftype) {
        case NL80211_IFTYPE_AP:
        case NL80211_IFTYPE_AP_VLAN:
+        case NL80211_IFTYPE_P2P_GO:
                /* disallow mesh-specific things */
                if (params.plink_action)
                        err = -EINVAL;
                break;
+        case NL80211_IFTYPE_P2P_CLIENT:
        case NL80211_IFTYPE_STATION:
                /* disallow everything but AUTHORIZED flag */
                if (params.plink_action)
@@ -2176,7 +2240,8 @@ static int nl80211_new_station(struct sk_buff *skb, struct genl_info *info)
                goto out_rtnl;
        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP &&
-            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP_VLAN) {
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP_VLAN &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_GO) {
                err = -EINVAL;
                goto out;
        }
@@ -2229,7 +2294,8 @@ static int nl80211_del_station(struct sk_buff *skb, struct genl_info *info)
        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP &&
            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP_VLAN &&
-            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_MESH_POINT) {
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_MESH_POINT &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_GO) {
                err = -EINVAL;
                goto out;
        }
@@ -2603,7 +2669,8 @@ static int nl80211_set_bss(struct sk_buff *skb, struct genl_info *info)
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_AP &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_GO) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -3306,6 +3373,7 @@ static int nl80211_send_bss(struct sk_buff *msg, u32 pid, u32 seq, int flags,
        }
        switch (wdev->iftype) {
+        case NL80211_IFTYPE_P2P_CLIENT:
        case NL80211_IFTYPE_STATION:
                if (intbss == wdev->current_bss)
                        NLA_PUT_U32(msg, NL80211_BSS_STATUS,
@@ -3572,12 +3640,28 @@ static int nl80211_authenticate(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto unlock_rtnl;
+        if (key.idx >= 0) {
+                int i;
+                bool ok = false;
+                for (i = 0; i < rdev->wiphy.n_cipher_suites; i++) {
+                        if (key.p.cipher == rdev->wiphy.cipher_suites[i]) {
+                                ok = true;
+                                break;
+                        }
+                }
+                if (!ok) {
+                        err = -EINVAL;
+                        goto out;
+                }
+        }
        if (!rdev->ops->auth) {
                err = -EOPNOTSUPP;
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -3624,7 +3708,8 @@ unlock_rtnl:
        return err;
 }
-static int nl80211_crypto_settings(struct genl_info *info,
+static int nl80211_crypto_settings(struct cfg80211_registered_device *rdev,
+                                   struct genl_info *info,
                                   struct cfg80211_crypto_settings *settings,
                                   int cipher_limit)
 {
@@ -3632,6 +3717,19 @@ static int nl80211_crypto_settings(struct genl_info *info,
        settings->control_port = info->attrs[NL80211_ATTR_CONTROL_PORT];
+        if (info->attrs[NL80211_ATTR_CONTROL_PORT_ETHERTYPE]) {
+                u16 proto;
+                proto = nla_get_u16(
+                        info->attrs[NL80211_ATTR_CONTROL_PORT_ETHERTYPE]);
+                settings->control_port_ethertype = cpu_to_be16(proto);
+                if (!(rdev->wiphy.flags & WIPHY_FLAG_CONTROL_PORT_PROTOCOL) &&
+                    proto != ETH_P_PAE)
+                        return -EINVAL;
+                if (info->attrs[NL80211_ATTR_CONTROL_PORT_NO_ENCRYPT])
+                        settings->control_port_no_encrypt = true;
+        } else
+                settings->control_port_ethertype = cpu_to_be16(ETH_P_PAE);
        if (info->attrs[NL80211_ATTR_CIPHER_SUITES_PAIRWISE]) {
                void *data;
                int len, i;
@@ -3718,7 +3816,8 @@ static int nl80211_associate(struct sk_buff *skb, struct genl_info *info)
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -3759,7 +3858,7 @@ static int nl80211_associate(struct sk_buff *skb, struct genl_info *info)
        if (info->attrs[NL80211_ATTR_PREV_BSSID])
                prev_bssid = nla_data(info->attrs[NL80211_ATTR_PREV_BSSID]);
-        err = nl80211_crypto_settings(info, &crypto, 1);
+        err = nl80211_crypto_settings(rdev, info, &crypto, 1);
        if (!err)
                err = cfg80211_mlme_assoc(rdev, dev, chan, bssid, prev_bssid,
                                          ssid, ssid_len, ie, ie_len, use_mfp,
@@ -3802,7 +3901,8 @@ static int nl80211_deauthenticate(struct sk_buff *skb, struct genl_info *info)
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -3868,7 +3968,8 @@ static int nl80211_disassociate(struct sk_buff *skb, struct genl_info *info)
                goto out;
        }
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4018,23 +4119,6 @@ static int nl80211_join_ibss(struct sk_buff *skb, struct genl_info *info)
                                goto out;
                        }
                }
-        } else {
-                /*
-                * If no rates were explicitly configured,
-                * use the mandatory rate set for 11b or
-                * 11a for maximum compatibility.
-                */
-                struct ieee80211_supported_band *sband =
-                        wiphy->bands[ibss.channel->band];
-                int j;
-                u32 flag = ibss.channel->band == IEEE80211_BAND_5GHZ ?
-                        IEEE80211_RATE_MANDATORY_A :
-                        IEEE80211_RATE_MANDATORY_B;
-                for (j = 0; j < sband->n_bitrates; j++) {
-                        if (sband->bitrates[j].flags & flag)
-                                ibss.basic_rates |= BIT(j);
-                }
        }
        err = cfg80211_join_ibss(rdev, dev, &ibss, connkeys);
@@ -4236,7 +4320,7 @@ static int nl80211_connect(struct sk_buff *skb, struct genl_info *info)
        connect.privacy = info->attrs[NL80211_ATTR_PRIVACY];
-        err = nl80211_crypto_settings(info, &connect.crypto,
+        err = nl80211_crypto_settings(rdev, info, &connect.crypto,
                                      NL80211_MAX_NR_CIPHER_SUITES);
        if (err)
                return err;
@@ -4246,7 +4330,8 @@ static int nl80211_connect(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto unlock_rtnl;
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4322,7 +4407,8 @@ static int nl80211_disconnect(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto unlock_rtnl;
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4410,7 +4496,8 @@ static int nl80211_setdel_pmksa(struct sk_buff *skb, struct genl_info *info)
        pmksa.pmkid = nla_data(info->attrs[NL80211_ATTR_PMKID]);
        pmksa.bssid = nla_data(info->attrs[NL80211_ATTR_MAC]);
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4455,7 +4542,8 @@ static int nl80211_flush_pmksa(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto out_rtnl;
-        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION) {
+        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4717,17 +4805,18 @@ static int nl80211_set_tx_bitrate_mask(struct sk_buff *skb,
        return err;
 }
-static int nl80211_register_action(struct sk_buff *skb, struct genl_info *info)
+static int nl80211_register_mgmt(struct sk_buff *skb, struct genl_info *info)
 {
        struct cfg80211_registered_device *rdev;
        struct net_device *dev;
+        u16 frame_type = IEEE80211_FTYPE_MGMT | IEEE80211_STYPE_ACTION;
        int err;
        if (!info->attrs[NL80211_ATTR_FRAME_MATCH])
                return -EINVAL;
-        if (nla_len(info->attrs[NL80211_ATTR_FRAME_MATCH]) < 1)
+        if (info->attrs[NL80211_ATTR_FRAME_TYPE])
-                return -EINVAL;
+                frame_type = nla_get_u16(info->attrs[NL80211_ATTR_FRAME_TYPE]);
        rtnl_lock();
@@ -4736,18 +4825,20 @@ static int nl80211_register_action(struct sk_buff *skb, struct genl_info *info)
                goto unlock_rtnl;
        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
-            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_ADHOC) {
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_ADHOC &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
        /* not much point in registering if we can't reply */
-        if (!rdev->ops->action) {
+        if (!rdev->ops->mgmt_tx) {
                err = -EOPNOTSUPP;
                goto out;
        }
-        err = cfg80211_mlme_register_action(dev->ieee80211_ptr, info->snd_pid,
+        err = cfg80211_mlme_register_mgmt(dev->ieee80211_ptr, info->snd_pid,
+                        frame_type,
                        nla_data(info->attrs[NL80211_ATTR_FRAME_MATCH]),
                        nla_len(info->attrs[NL80211_ATTR_FRAME_MATCH]));
 out:
@@ -4758,7 +4849,7 @@ static int nl80211_register_action(struct sk_buff *skb, struct genl_info *info)
        return err;
 }
-static int nl80211_action(struct sk_buff *skb, struct genl_info *info)
+static int nl80211_tx_mgmt(struct sk_buff *skb, struct genl_info *info)
 {
        struct cfg80211_registered_device *rdev;
        struct net_device *dev;
@@ -4781,13 +4872,14 @@ static int nl80211_action(struct sk_buff *skb, struct genl_info *info)
        if (err)
                goto unlock_rtnl;
-        if (!rdev->ops->action) {
+        if (!rdev->ops->mgmt_tx) {
                err = -EOPNOTSUPP;
                goto out;
        }
        if (dev->ieee80211_ptr->iftype != NL80211_IFTYPE_STATION &&
-            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_ADHOC) {
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_ADHOC &&
+            dev->ieee80211_ptr->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto out;
        }
@@ -4824,17 +4916,17 @@ static int nl80211_action(struct sk_buff *skb, struct genl_info *info)
        }
        hdr = nl80211hdr_put(msg, info->snd_pid, info->snd_seq, 0,
-                             NL80211_CMD_ACTION);
+                             NL80211_CMD_FRAME);
        if (IS_ERR(hdr)) {
                err = PTR_ERR(hdr);
                goto free_msg;
        }
-        err = cfg80211_mlme_action(rdev, dev, chan, channel_type,
+        err = cfg80211_mlme_mgmt_tx(rdev, dev, chan, channel_type,
-                                   channel_type_valid,
+                                    channel_type_valid,
-                                   nla_data(info->attrs[NL80211_ATTR_FRAME]),
+                                    nla_data(info->attrs[NL80211_ATTR_FRAME]),
-                                   nla_len(info->attrs[NL80211_ATTR_FRAME]),
+                                    nla_len(info->attrs[NL80211_ATTR_FRAME]),
-                                   &cookie);
+                                    &cookie);
        if (err)
                goto free_msg;
@@ -4881,7 +4973,7 @@ static int nl80211_set_power_save(struct sk_buff *skb, struct genl_info *info)
        err = get_rdev_dev_by_info_ifindex(info, &rdev, &dev);
        if (err)
-                goto unlock_rdev;
+                goto unlock_rtnl;
        wdev = dev->ieee80211_ptr;
@@ -4905,6 +4997,7 @@ static int nl80211_set_power_save(struct sk_buff *skb, struct genl_info *info)
 unlock_rdev:
        cfg80211_unlock_rdev(rdev);
        dev_put(dev);
+unlock_rtnl:
        rtnl_unlock();
 out:
@@ -5005,7 +5098,8 @@ static int nl80211_set_cqm_rssi(struct genl_info *info,
                goto unlock_rdev;
        }
-        if (wdev->iftype != NL80211_IFTYPE_STATION) {
+        if (wdev->iftype != NL80211_IFTYPE_STATION &&
+            wdev->iftype != NL80211_IFTYPE_P2P_CLIENT) {
                err = -EOPNOTSUPP;
                goto unlock_rdev;
        }
@@ -5333,14 +5427,14 @@ static struct genl_ops nl80211_ops[] = {
                .flags = GENL_ADMIN_PERM,
        },
        {
-                .cmd = NL80211_CMD_REGISTER_ACTION,
+                .cmd = NL80211_CMD_REGISTER_FRAME,
-                .doit = nl80211_register_action,
+                .doit = nl80211_register_mgmt,
                .policy = nl80211_policy,
                .flags = GENL_ADMIN_PERM,
        },
        {
-                .cmd = NL80211_CMD_ACTION,
+                .cmd = NL80211_CMD_FRAME,
-                .doit = nl80211_action,
+                .doit = nl80211_tx_mgmt,
                .policy = nl80211_policy,
                .flags = GENL_ADMIN_PERM,
        },
@@ -6040,9 +6134,9 @@ void nl80211_send_sta_event(struct cfg80211_registered_device *rdev,
                                nl80211_mlme_mcgrp.id, gfp);
 }
-int nl80211_send_action(struct cfg80211_registered_device *rdev,
+int nl80211_send_mgmt(struct cfg80211_registered_device *rdev,
-                        struct net_device *netdev, u32 nlpid,
+                      struct net_device *netdev, u32 nlpid,
-                        int freq, const u8 *buf, size_t len, gfp_t gfp)
+                      int freq, const u8 *buf, size_t len, gfp_t gfp)
 {
        struct sk_buff *msg;
        void *hdr;
@@ -6052,7 +6146,7 @@ int nl80211_send_action(struct cfg80211_registered_device *rdev,
        if (!msg)
                return -ENOMEM;
-        hdr = nl80211hdr_put(msg, 0, 0, 0, NL80211_CMD_ACTION);
+        hdr = nl80211hdr_put(msg, 0, 0, 0, NL80211_CMD_FRAME);
        if (!hdr) {
                nlmsg_free(msg);
                return -ENOMEM;
@@ -6080,10 +6174,10 @@ int nl80211_send_action(struct cfg80211_registered_device *rdev,
        return -ENOBUFS;
 }
-void nl80211_send_action_tx_status(struct cfg80211_registered_device *rdev,
+void nl80211_send_mgmt_tx_status(struct cfg80211_registered_device *rdev,
-                                   struct net_device *netdev, u64 cookie,
+                                 struct net_device *netdev, u64 cookie,
-                                   const u8 *buf, size_t len, bool ack,
+                                 const u8 *buf, size_t len, bool ack,
-                                   gfp_t gfp)
+                                 gfp_t gfp)
 {
        struct sk_buff *msg;
        void *hdr;
@@ -6092,7 +6186,7 @@ void nl80211_send_action_tx_status(struct cfg80211_registered_device *rdev,
        if (!msg)
                return;
-        hdr = nl80211hdr_put(msg, 0, 0, 0, NL80211_CMD_ACTION_TX_STATUS);
+        hdr = nl80211hdr_put(msg, 0, 0, 0, NL80211_CMD_FRAME_TX_STATUS);
        if (!hdr) {
                nlmsg_free(msg);
                return;
@@ -6179,7 +6273,7 @@ static int nl80211_netlink_notify(struct notifier_block * nb,
        list_for_each_entry_rcu(rdev, &cfg80211_rdev_list, list)
                list_for_each_entry_rcu(wdev, &rdev->netdev_list, list)
-                        cfg80211_mlme_unregister_actions(wdev, notify->pid);
+                        cfg80211_mlme_unregister_socket(wdev, notify->pid);
        rcu_read_unlock();
diff --git a/net/wireless/nl80211.h b/net/wireless/nl80211.h
index 2ad7fbc7d9f1..30d2f939150d 100644
--- a/net/wireless/nl80211.h
+++ b/net/wireless/nl80211.h
@@ -74,13 +74,13 @@ void nl80211_send_sta_event(struct cfg80211_registered_device *rdev,
                            struct net_device *dev, const u8 *mac_addr,
                            struct station_info *sinfo, gfp_t gfp);
-int nl80211_send_action(struct cfg80211_registered_device *rdev,
+int nl80211_send_mgmt(struct cfg80211_registered_device *rdev,
-                        struct net_device *netdev, u32 nlpid, int freq,
+                      struct net_device *netdev, u32 nlpid, int freq,
-                        const u8 *buf, size_t len, gfp_t gfp);
+                      const u8 *buf, size_t len, gfp_t gfp);
-void nl80211_send_action_tx_status(struct cfg80211_registered_device *rdev,
+void nl80211_send_mgmt_tx_status(struct cfg80211_registered_device *rdev,
-                                   struct net_device *netdev, u64 cookie,
+                                 struct net_device *netdev, u64 cookie,
-                                   const u8 *buf, size_t len, bool ack,
+                                 const u8 *buf, size_t len, bool ack,
-                                   gfp_t gfp);
+                                 gfp_t gfp);
 void
 nl80211_send_cqm_rssi_notify(struct cfg80211_registered_device *rdev,
diff --git a/net/wireless/radiotap.c b/net/wireless/radiotap.c
index 1332c445d1c7..c774bc0f155e 100644
--- a/net/wireless/radiotap.c
+++ b/net/wireless/radiotap.c
@@ -14,6 +14,7 @@
 * See COPYING for more details.
 */
+#include <linux/kernel.h>
 #include <net/cfg80211.h>
 #include <net/ieee80211_radiotap.h>
 #include <asm/unaligned.h>
@@ -45,7 +46,7 @@ static const struct radiotap_align_size rtap_namespace_sizes[] = {
 };
 static const struct ieee80211_radiotap_namespace radiotap_ns = {
-        .n_bits = sizeof(rtap_namespace_sizes) / sizeof(rtap_namespace_sizes[0]),
+        .n_bits = ARRAY_SIZE(rtap_namespace_sizes),
        .align_size = rtap_namespace_sizes,
 };
diff --git a/net/wireless/reg.c b/net/wireless/reg.c
index f180db0de66c..d14bbf960c18 100644
--- a/net/wireless/reg.c
+++ b/net/wireless/reg.c
@@ -36,6 +36,7 @@
 #include <linux/slab.h>
 #include <linux/list.h>
 #include <linux/random.h>
+#include <linux/ctype.h>
 #include <linux/nl80211.h>
 #include <linux/platform_device.h>
 #include <net/cfg80211.h>
@@ -73,7 +74,11 @@ const struct ieee80211_regdomain *cfg80211_regdomain;
 *     - last_request
 */
 static DEFINE_MUTEX(reg_mutex);
-#define assert_reg_lock() WARN_ON(!mutex_is_locked(&reg_mutex))
+static inline void assert_reg_lock(void)
+{
+        lockdep_assert_held(&reg_mutex);
+}
 /* Used to queue up regulatory hints */
 static LIST_HEAD(reg_requests_list);
@@ -181,14 +186,6 @@ static bool is_alpha2_set(const char *alpha2)
        return false;
 }
-static bool is_alpha_upper(char letter)
-{
-        /* ASCII A - Z */
-        if (letter >= 65 && letter <= 90)
-                return true;
-        return false;
-}
 static bool is_unknown_alpha2(const char *alpha2)
 {
        if (!alpha2)
@@ -220,7 +217,7 @@ static bool is_an_alpha2(const char *alpha2)
 {
        if (!alpha2)
                return false;
-        if (is_alpha_upper(alpha2[0]) && is_alpha_upper(alpha2[1]))
+        if (isalpha(alpha2[0]) && isalpha(alpha2[1]))
                return true;
        return false;
 }
@@ -1399,6 +1396,11 @@ static DECLARE_WORK(reg_work, reg_todo);
 static void queue_regulatory_request(struct regulatory_request *request)
 {
+        if (isalpha(request->alpha2[0]))
+                request->alpha2[0] = toupper(request->alpha2[0]);
+        if (isalpha(request->alpha2[1]))
+                request->alpha2[1] = toupper(request->alpha2[1]);
        spin_lock(&reg_requests_lock);
        list_add_tail(&request->list, &reg_requests_list);
        spin_unlock(&reg_requests_lock);
diff --git a/net/wireless/sme.c b/net/wireless/sme.c
index a8c2d6b877ae..f161b9844542 100644
--- a/net/wireless/sme.c
+++ b/net/wireless/sme.c
@@ -411,7 +411,8 @@ void __cfg80211_connect_result(struct net_device *dev, const u8 *bssid,
        ASSERT_WDEV_LOCK(wdev);
-        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION))
+        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION &&
+                    wdev->iftype != NL80211_IFTYPE_P2P_CLIENT))
                return;
        if (wdev->sme_state != CFG80211_SME_CONNECTING)
@@ -548,7 +549,8 @@ void __cfg80211_roamed(struct wireless_dev *wdev, const u8 *bssid,
        ASSERT_WDEV_LOCK(wdev);
-        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION))
+        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION &&
+                    wdev->iftype != NL80211_IFTYPE_P2P_CLIENT))
                return;
        if (wdev->sme_state != CFG80211_SME_CONNECTED)
@@ -644,7 +646,8 @@ void __cfg80211_disconnected(struct net_device *dev, const u8 *ie,
        ASSERT_WDEV_LOCK(wdev);
-        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION))
+        if (WARN_ON(wdev->iftype != NL80211_IFTYPE_STATION &&
+                    wdev->iftype != NL80211_IFTYPE_P2P_CLIENT))
                return;
        if (wdev->sme_state != CFG80211_SME_CONNECTED)
diff --git a/net/wireless/sysfs.c b/net/wireless/sysfs.c
index 9f2cef3e0ca0..74a9e3cce452 100644
--- a/net/wireless/sysfs.c
+++ b/net/wireless/sysfs.c
@@ -110,6 +110,13 @@ static int wiphy_resume(struct device *dev)
        return ret;
 }
+static const void *wiphy_namespace(struct device *d)
+{
+        struct wiphy *wiphy = container_of(d, struct wiphy, dev);
+        return wiphy_net(wiphy);
+}
 struct class ieee80211_class = {
        .name = "ieee80211",
        .owner = THIS_MODULE,
@@ -120,6 +127,8 @@ struct class ieee80211_class = {
 #endif
        .suspend = wiphy_suspend,
        .resume = wiphy_resume,
+        .ns_type = &net_ns_type_operations,
+        .namespace = wiphy_namespace,
 };
 int wiphy_sysfs_init(void)
diff --git a/net/wireless/util.c b/net/wireless/util.c
index 0c8a1e8b7690..fb5448f7d55a 100644
--- a/net/wireless/util.c
+++ b/net/wireless/util.c
@@ -183,7 +183,14 @@ int cfg80211_validate_key_settings(struct cfg80211_registered_device *rdev,
                        return -EINVAL;
                break;
        default:
-                return -EINVAL;
+                /*
+                 * We don't know anything about this algorithm,
+                 * allow using it -- but the driver must check
+                 * all parameters! We still check below whether
+                 * or not the driver supports this algorithm,
+                 * of course.
+                 */
+                break;
        }
        if (params->seq) {
@@ -221,7 +228,7 @@ const unsigned char bridge_tunnel_header[] __aligned(2) =
        { 0xaa, 0xaa, 0x03, 0x00, 0x00, 0xf8 };
 EXPORT_SYMBOL(bridge_tunnel_header);
-unsigned int ieee80211_hdrlen(__le16 fc)
+unsigned int __attribute_const__ ieee80211_hdrlen(__le16 fc)
 {
        unsigned int hdrlen = 24;
@@ -319,7 +326,8 @@ int ieee80211_data_to_8023(struct sk_buff *skb, const u8 *addr,
                cpu_to_le16(IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS)) {
        case cpu_to_le16(IEEE80211_FCTL_TODS):
                if (unlikely(iftype != NL80211_IFTYPE_AP &&
-                             iftype != NL80211_IFTYPE_AP_VLAN))
+                             iftype != NL80211_IFTYPE_AP_VLAN &&
+                             iftype != NL80211_IFTYPE_P2P_GO))
                        return -1;
                break;
        case cpu_to_le16(IEEE80211_FCTL_TODS | IEEE80211_FCTL_FROMDS):
@@ -347,7 +355,8 @@ int ieee80211_data_to_8023(struct sk_buff *skb, const u8 *addr,
                break;
        case cpu_to_le16(IEEE80211_FCTL_FROMDS):
                if ((iftype != NL80211_IFTYPE_STATION &&
-                    iftype != NL80211_IFTYPE_MESH_POINT) ||
+                     iftype != NL80211_IFTYPE_P2P_CLIENT &&
+                     iftype != NL80211_IFTYPE_MESH_POINT) ||
                    (is_multicast_ether_addr(dst) &&
                     !compare_ether_addr(src, addr)))
                        return -1;
@@ -424,6 +433,7 @@ int ieee80211_data_from_8023(struct sk_buff *skb, const u8 *addr,
        switch (iftype) {
        case NL80211_IFTYPE_AP:
        case NL80211_IFTYPE_AP_VLAN:
+        case NL80211_IFTYPE_P2P_GO:
                fc |= cpu_to_le16(IEEE80211_FCTL_FROMDS);
                /* DA BSSID SA */
                memcpy(hdr.addr1, skb->data, ETH_ALEN);
@@ -432,6 +442,7 @@ int ieee80211_data_from_8023(struct sk_buff *skb, const u8 *addr,
                hdrlen = 24;
                break;
        case NL80211_IFTYPE_STATION:
+        case NL80211_IFTYPE_P2P_CLIENT:
                fc |= cpu_to_le16(IEEE80211_FCTL_TODS);
                /* BSSID SA DA */
                memcpy(hdr.addr1, bssid, ETH_ALEN);
@@ -771,7 +782,9 @@ int cfg80211_change_iface(struct cfg80211_registered_device *rdev,
        /* if it's part of a bridge, reject changing type to station/ibss */
        if ((dev->priv_flags & IFF_BRIDGE_PORT) &&
-            (ntype == NL80211_IFTYPE_ADHOC || ntype == NL80211_IFTYPE_STATION))
+            (ntype == NL80211_IFTYPE_ADHOC ||
+             ntype == NL80211_IFTYPE_STATION ||
+             ntype == NL80211_IFTYPE_P2P_CLIENT))
                return -EBUSY;
        if (ntype != otype) {
@@ -782,6 +795,7 @@ int cfg80211_change_iface(struct cfg80211_registered_device *rdev,
                        cfg80211_leave_ibss(rdev, dev, false);
                        break;
                case NL80211_IFTYPE_STATION:
+                case NL80211_IFTYPE_P2P_CLIENT:
                        cfg80211_disconnect(rdev, dev,
                                            WLAN_REASON_DEAUTH_LEAVING, true);
                        break;
@@ -810,9 +824,11 @@ int cfg80211_change_iface(struct cfg80211_registered_device *rdev,
                        if (dev->ieee80211_ptr->use_4addr)
                                break;
                        /* fall through */
+                case NL80211_IFTYPE_P2P_CLIENT:
                case NL80211_IFTYPE_ADHOC:
                        dev->priv_flags |= IFF_DONT_BRIDGE;
                        break;
+                case NL80211_IFTYPE_P2P_GO:
                case NL80211_IFTYPE_AP:
                case NL80211_IFTYPE_AP_VLAN:
                case NL80211_IFTYPE_WDS:
@@ -823,7 +839,7 @@ int cfg80211_change_iface(struct cfg80211_registered_device *rdev,
                        /* monitor can't bridge anyway */
                        break;
                case NL80211_IFTYPE_UNSPECIFIED:
-                case __NL80211_IFTYPE_AFTER_LAST:
+                case NUM_NL80211_IFTYPES:
                        /* not happening */
                        break;
                }
diff --git a/net/wireless/wext-core.c b/net/wireless/wext-core.c
index 8f5116f5af19..dc675a3daa3d 100644
--- a/net/wireless/wext-core.c
+++ b/net/wireless/wext-core.c
@@ -611,7 +611,7 @@ struct iw_statistics *get_wireless_stats(struct net_device *dev)
 #endif
 #ifdef CONFIG_CFG80211_WEXT
-        if (dev->ieee80211_ptr && dev->ieee80211_ptr &&
+        if (dev->ieee80211_ptr &&
            dev->ieee80211_ptr->wiphy &&
            dev->ieee80211_ptr->wiphy->wext &&
            dev->ieee80211_ptr->wiphy->wext->get_wireless_stats)
diff --git a/net/wireless/wext-sme.c b/net/wireless/wext-sme.c
index 9818198add8a..6fffe62d7c25 100644
--- a/net/wireless/wext-sme.c
+++ b/net/wireless/wext-sme.c
@@ -197,6 +197,8 @@ int cfg80211_mgd_wext_siwessid(struct net_device *dev,
        wdev->wext.connect.ssid_len = len;
        wdev->wext.connect.crypto.control_port = false;
+        wdev->wext.connect.crypto.control_port_ethertype =
+                                        cpu_to_be16(ETH_P_PAE);
        err = cfg80211_mgd_wext_connect(rdev, wdev);
 out:
diff --git a/net/x25/af_x25.c b/net/x25/af_x25.c
index 5e86d4e97dce..f7af98dff409 100644
--- a/net/x25/af_x25.c
+++ b/net/x25/af_x25.c
@@ -507,14 +507,14 @@ static int x25_listen(struct socket *sock, int backlog)
        struct sock *sk = sock->sk;
        int rc = -EOPNOTSUPP;
-        lock_kernel();
+        lock_sock(sk);
        if (sk->sk_state != TCP_LISTEN) {
                memset(&x25_sk(sk)->dest_addr, 0, X25_ADDR_LEN);
                sk->sk_max_ack_backlog = backlog;
                sk->sk_state           = TCP_LISTEN;
                rc = 0;
        }
-        unlock_kernel();
+        release_sock(sk);
        return rc;
 }
@@ -688,7 +688,6 @@ static int x25_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
        struct sockaddr_x25 *addr = (struct sockaddr_x25 *)uaddr;
        int len, i, rc = 0;
-        lock_kernel();
        if (!sock_flag(sk, SOCK_ZAPPED) ||
            addr_len != sizeof(struct sockaddr_x25) ||
            addr->sx25_family != AF_X25) {
@@ -704,12 +703,13 @@ static int x25_bind(struct socket *sock, struct sockaddr *uaddr, int addr_len)
                }
        }
+        lock_sock(sk);
        x25_sk(sk)->source_addr = addr->sx25_addr;
        x25_insert_socket(sk);
        sock_reset_flag(sk, SOCK_ZAPPED);
+        release_sock(sk);
        SOCK_DEBUG(sk, "x25_bind: socket is bound\n");
 out:
-        unlock_kernel();
        return rc;
 }
@@ -751,7 +751,6 @@ static int x25_connect(struct socket *sock, struct sockaddr *uaddr,
        struct x25_route *rt;
        int rc = 0;
-        lock_kernel();
        lock_sock(sk);
        if (sk->sk_state == TCP_ESTABLISHED && sock->state == SS_CONNECTING) {
                sock->state = SS_CONNECTED;
@@ -829,7 +828,6 @@ out_put_route:
        x25_route_put(rt);
 out:
        release_sock(sk);
-        unlock_kernel();
        return rc;
 }
@@ -869,8 +867,7 @@ static int x25_accept(struct socket *sock, struct socket *newsock, int flags)
        struct sk_buff *skb;
        int rc = -EINVAL;
-        lock_kernel();
+        if (!sk)
-        if (!sk || sk->sk_state != TCP_LISTEN)
                goto out;
        rc = -EOPNOTSUPP;
@@ -878,6 +875,10 @@ static int x25_accept(struct socket *sock, struct socket *newsock, int flags)
                goto out;
        lock_sock(sk);
+        rc = -EINVAL;
+        if (sk->sk_state != TCP_LISTEN)
+                goto out2;
        rc = x25_wait_for_data(sk, sk->sk_rcvtimeo);
        if (rc)
                goto out2;
@@ -897,7 +898,6 @@ static int x25_accept(struct socket *sock, struct socket *newsock, int flags)
 out2:
        release_sock(sk);
 out:
-        unlock_kernel();
        return rc;
 }
@@ -909,7 +909,6 @@ static int x25_getname(struct socket *sock, struct sockaddr *uaddr,
        struct x25_sock *x25 = x25_sk(sk);
        int rc = 0;
-        lock_kernel();
        if (peer) {
                if (sk->sk_state != TCP_ESTABLISHED) {
                        rc = -ENOTCONN;
@@ -923,19 +922,6 @@ static int x25_getname(struct socket *sock, struct sockaddr *uaddr,
        *uaddr_len = sizeof(*sx25);
 out:
-        unlock_kernel();
-        return rc;
-}
-static unsigned int x25_datagram_poll(struct file *file, struct socket *sock,
-                           poll_table *wait)
-{
-        int rc;
-        lock_kernel();
-        rc = datagram_poll(file, sock, wait);
-        unlock_kernel();
        return rc;
 }
@@ -1746,7 +1732,7 @@ static const struct proto_ops x25_proto_ops = {
        .socketpair =   sock_no_socketpair,
        .accept =       x25_accept,
        .getname =      x25_getname,
-        .poll =         x25_datagram_poll,
+        .poll =         datagram_poll,
        .ioctl =        x25_ioctl,
 #ifdef CONFIG_COMPAT
        .compat_ioctl = compat_x25_ioctl,