10 files changed, 311 insertions, 23 deletions

diff --git a/include/rdma/Kbuild b/include/rdma/Kbuild
index e7c043216558..ea56f76c0c22 100644
--- a/include/rdma/Kbuild
+++ b/include/rdma/Kbuild
@@ -1 +1,6 @@
+header-y += ib_user_cm.h
 header-y += ib_user_mad.h
+header-y += ib_user_sa.h
+header-y += ib_user_verbs.h
+header-y += rdma_netlink.h
+header-y += rdma_user_cm.h
diff --git a/include/rdma/ib_addr.h b/include/rdma/ib_addr.h
index fa0d52b8e622..ae8c68f30f1b 100644
--- a/include/rdma/ib_addr.h
+++ b/include/rdma/ib_addr.h
@@ -39,7 +39,9 @@
 #include <linux/if_arp.h>
 #include <linux/netdevice.h>
 #include <linux/socket.h>
+#include <linux/if_vlan.h>
 #include <rdma/ib_verbs.h>
+#include <rdma/ib_pack.h>
 
 struct rdma_addr_client {
         atomic_t refcount;
@@ -63,6 +65,7 @@ struct rdma_dev_addr {
         unsigned char broadcast[MAX_ADDR_LEN];
         unsigned short dev_type;
         int bound_dev_if;
+        enum rdma_transport_type transport;
 };
 
 /**
@@ -127,9 +130,51 @@ static inline int rdma_addr_gid_offset(struct rdma_dev_addr *dev_addr)
         return dev_addr->dev_type == ARPHRD_INFINIBAND ? 4 : 0;
 }
 
+static inline void iboe_mac_vlan_to_ll(union ib_gid *gid, u8 *mac, u16 vid)
+{
+        memset(gid->raw, 0, 16);
+        *((__be32 *) gid->raw) = cpu_to_be32(0xfe800000);
+        if (vid < 0x1000) {
+                gid->raw[12] = vid & 0xff;
+                gid->raw[11] = vid >> 8;
+        } else {
+                gid->raw[12] = 0xfe;
+                gid->raw[11] = 0xff;
+        }
+        memcpy(gid->raw + 13, mac + 3, 3);
+        memcpy(gid->raw + 8, mac, 3);
+        gid->raw[8] ^= 2;
+}
+
+static inline u16 rdma_vlan_dev_vlan_id(const struct net_device *dev)
+{
+        return dev->priv_flags & IFF_802_1Q_VLAN ?
+                vlan_dev_vlan_id(dev) : 0xffff;
+}
+
+static inline void iboe_addr_get_sgid(struct rdma_dev_addr *dev_addr,
+                                      union ib_gid *gid)
+{
+        struct net_device *dev;
+        u16 vid = 0xffff;
+
+        dev = dev_get_by_index(&init_net, dev_addr->bound_dev_if);
+        if (dev) {
+                vid = rdma_vlan_dev_vlan_id(dev);
+                dev_put(dev);
+        }
+
+        iboe_mac_vlan_to_ll(gid, dev_addr->src_dev_addr, vid);
+}
+
 static inline void rdma_addr_get_sgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
 {
-        memcpy(gid, dev_addr->src_dev_addr + rdma_addr_gid_offset(dev_addr), sizeof *gid);
+        if (dev_addr->transport == RDMA_TRANSPORT_IB &&
+            dev_addr->dev_type != ARPHRD_INFINIBAND)
+                iboe_addr_get_sgid(dev_addr, gid);
+        else
+                memcpy(gid, dev_addr->src_dev_addr +
+                       rdma_addr_gid_offset(dev_addr), sizeof *gid);
 }
 
 static inline void rdma_addr_set_sgid(struct rdma_dev_addr *dev_addr, union ib_gid *gid)
@@ -147,4 +192,92 @@ static inline void rdma_addr_set_dgid(struct rdma_dev_addr *dev_addr, union ib_g
         memcpy(dev_addr->dst_dev_addr + rdma_addr_gid_offset(dev_addr), gid, sizeof *gid);
 }
 
+static inline enum ib_mtu iboe_get_mtu(int mtu)
+{
+        /*
+         * reduce IB headers from effective IBoE MTU. 28 stands for
+         * atomic header which is the biggest possible header after BTH
+         */
+        mtu = mtu - IB_GRH_BYTES - IB_BTH_BYTES - 28;
+
+        if (mtu >= ib_mtu_enum_to_int(IB_MTU_4096))
+                return IB_MTU_4096;
+        else if (mtu >= ib_mtu_enum_to_int(IB_MTU_2048))
+                return IB_MTU_2048;
+        else if (mtu >= ib_mtu_enum_to_int(IB_MTU_1024))
+                return IB_MTU_1024;
+        else if (mtu >= ib_mtu_enum_to_int(IB_MTU_512))
+                return IB_MTU_512;
+        else if (mtu >= ib_mtu_enum_to_int(IB_MTU_256))
+                return IB_MTU_256;
+        else
+                return 0;
+}
+
+static inline int iboe_get_rate(struct net_device *dev)
+{
+        struct ethtool_cmd cmd;
+        u32 speed;
+
+        if (dev_ethtool_get_settings(dev, &cmd))
+                return IB_RATE_PORT_CURRENT;
+
+        speed = ethtool_cmd_speed(&cmd);
+        if (speed >= 40000)
+                return IB_RATE_40_GBPS;
+        else if (speed >= 30000)
+                return IB_RATE_30_GBPS;
+        else if (speed >= 20000)
+                return IB_RATE_20_GBPS;
+        else if (speed >= 10000)
+                return IB_RATE_10_GBPS;
+        else
+                return IB_RATE_PORT_CURRENT;
+}
+
+static inline int rdma_link_local_addr(struct in6_addr *addr)
+{
+        if (addr->s6_addr32[0] == htonl(0xfe800000) &&
+            addr->s6_addr32[1] == 0)
+                return 1;
+
+        return 0;
+}
+
+static inline void rdma_get_ll_mac(struct in6_addr *addr, u8 *mac)
+{
+        memcpy(mac, &addr->s6_addr[8], 3);
+        memcpy(mac + 3, &addr->s6_addr[13], 3);
+        mac[0] ^= 2;
+}
+
+static inline int rdma_is_multicast_addr(struct in6_addr *addr)
+{
+        return addr->s6_addr[0] == 0xff;
+}
+
+static inline void rdma_get_mcast_mac(struct in6_addr *addr, u8 *mac)
+{
+        int i;
+
+        mac[0] = 0x33;
+        mac[1] = 0x33;
+        for (i = 2; i < 6; ++i)
+                mac[i] = addr->s6_addr[i + 10];
+}
+
+static inline u16 rdma_get_vlan_id(union ib_gid *dgid)
+{
+        u16 vid;
+
+        vid = dgid->raw[11] << 8 | dgid->raw[12];
+        return vid < 0x1000 ? vid : 0xffff;
+}
+
+static inline struct net_device *rdma_vlan_dev_real_dev(const struct net_device *dev)
+{
+        return dev->priv_flags & IFF_802_1Q_VLAN ?
+                vlan_dev_real_dev(dev) : 0;
+}
+
 #endif /* IB_ADDR_H */
diff --git a/include/rdma/ib_pack.h b/include/rdma/ib_pack.h
index cbb50f4da3dd..b37fe3b10a9d 100644
--- a/include/rdma/ib_pack.h
+++ b/include/rdma/ib_pack.h
@@ -37,6 +37,8 @@
 
 enum {
         IB_LRH_BYTES  = 8,
+        IB_ETH_BYTES  = 14,
+        IB_VLAN_BYTES = 4,
         IB_GRH_BYTES  = 40,
         IB_BTH_BYTES  = 12,
         IB_DETH_BYTES = 8
@@ -210,14 +212,32 @@ struct ib_unpacked_deth {
         __be32       source_qpn;
 };
 
+struct ib_unpacked_eth {
+        u8      dmac_h[4];
+        u8      dmac_l[2];
+        u8      smac_h[2];
+        u8      smac_l[4];
+        __be16  type;
+};
+
+struct ib_unpacked_vlan {
+        __be16  tag;
+        __be16  type;
+};
+
 struct ib_ud_header {
+        int                     lrh_present;
         struct ib_unpacked_lrh  lrh;
-        int                     grh_present;
-        struct ib_unpacked_grh  grh;
-        struct ib_unpacked_bth  bth;
+        int                     eth_present;
+        struct ib_unpacked_eth  eth;
+        int                     vlan_present;
+        struct ib_unpacked_vlan vlan;
+        int                     grh_present;
+        struct ib_unpacked_grh  grh;
+        struct ib_unpacked_bth  bth;
         struct ib_unpacked_deth deth;
-        int                     immediate_present;
-        __be32                  immediate_data;
+        int                     immediate_present;
+        __be32                  immediate_data;
 };
 
 void ib_pack(const struct ib_field        *desc,
@@ -230,9 +250,12 @@ void ib_unpack(const struct ib_field        *desc,
                void                         *buf,
                void                         *structure);
 
-void ib_ud_header_init(int                 payload_bytes,
-                       int                 grh_present,
-                       int                 immediate_present,
+void ib_ud_header_init(int                  payload_bytes,
+                       int                  lrh_present,
+                       int                  eth_present,
+                       int                  vlan_present,
+                       int                  grh_present,
+                       int                  immediate_present,
                        struct ib_ud_header *header);
 
 int ib_ud_header_pack(struct ib_ud_header *header,
diff --git a/include/rdma/ib_user_cm.h b/include/rdma/ib_user_cm.h
index bd3d380781e0..f79014aa28f9 100644
--- a/include/rdma/ib_user_cm.h
+++ b/include/rdma/ib_user_cm.h
@@ -34,6 +34,7 @@
 #ifndef IB_USER_CM_H
 #define IB_USER_CM_H
 
+#include <linux/types.h>
 #include <rdma/ib_user_sa.h>
 
 #define IB_USER_CM_ABI_VERSION 5
diff --git a/include/rdma/ib_user_verbs.h b/include/rdma/ib_user_verbs.h
index a17f77106149..fe5b05177a2c 100644
--- a/include/rdma/ib_user_verbs.h
+++ b/include/rdma/ib_user_verbs.h
@@ -205,7 +205,8 @@ struct ib_uverbs_query_port_resp {
         __u8  active_width;
         __u8  active_speed;
         __u8  phys_state;
-        __u8  reserved[3];
+        __u8  link_layer;
+        __u8  reserved[2];
 };
 
 struct ib_uverbs_alloc_pd {
diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 857b3b9cf120..55cd0a0bc977 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -47,10 +47,13 @@
 #include <linux/list.h>
 #include <linux/rwsem.h>
 #include <linux/scatterlist.h>
+#include <linux/workqueue.h>
 
 #include <asm/atomic.h>
 #include <asm/uaccess.h>
 
+extern struct workqueue_struct *ib_wq;
+
 union ib_gid {
         u8      raw[16];
         struct {
@@ -75,6 +78,12 @@ enum rdma_transport_type {
 enum rdma_transport_type
 rdma_node_get_transport(enum rdma_node_type node_type) __attribute_const__;
 
+enum rdma_link_layer {
+        IB_LINK_LAYER_UNSPECIFIED,
+        IB_LINK_LAYER_INFINIBAND,
+        IB_LINK_LAYER_ETHERNET,
+};
+
 enum ib_device_cap_flags {
         IB_DEVICE_RESIZE_MAX_WR         = 1,
         IB_DEVICE_BAD_PKEY_CNTR         = (1<<1),
@@ -1010,6 +1019,8 @@ struct ib_device {
         int                        (*query_port)(struct ib_device *device,
                                                  u8 port_num,
                                                  struct ib_port_attr *port_attr);
+        enum rdma_link_layer       (*get_link_layer)(struct ib_device *device,
+                                                     u8 port_num);
         int                        (*query_gid)(struct ib_device *device,
                                                 u8 port_num, int index,
                                                 union ib_gid *gid);
@@ -1222,6 +1233,9 @@ int ib_query_device(struct ib_device *device,
 int ib_query_port(struct ib_device *device,
                   u8 port_num, struct ib_port_attr *port_attr);
 
+enum rdma_link_layer rdma_port_get_link_layer(struct ib_device *device,
+                                               u8 port_num);
+
 int ib_query_gid(struct ib_device *device,
                  u8 port_num, int index, union ib_gid *gid);
 
diff --git a/include/rdma/iw_cm.h b/include/rdma/iw_cm.h
index cbb822e8d791..2d0191c90f9e 100644
--- a/include/rdma/iw_cm.h
+++ b/include/rdma/iw_cm.h
@@ -46,18 +46,9 @@ enum iw_cm_event_type {
         IW_CM_EVENT_CLOSE                /* close complete */
 };
 
-enum iw_cm_event_status {
-        IW_CM_EVENT_STATUS_OK = 0,       /* request successful */
-        IW_CM_EVENT_STATUS_ACCEPTED = 0, /* connect request accepted */
-        IW_CM_EVENT_STATUS_REJECTED,     /* connect request rejected */
-        IW_CM_EVENT_STATUS_TIMEOUT,      /* the operation timed out */
-        IW_CM_EVENT_STATUS_RESET,        /* reset from remote peer */
-        IW_CM_EVENT_STATUS_EINVAL,       /* asynchronous failure for bad parm */
-};
-
 struct iw_cm_event {
         enum iw_cm_event_type event;
-        enum iw_cm_event_status status;
+        int                      status;
         struct sockaddr_in local_addr;
         struct sockaddr_in remote_addr;
         void *private_data;
diff --git a/include/rdma/rdma_cm.h b/include/rdma/rdma_cm.h
index 4fae90304648..26977c149c41 100644
--- a/include/rdma/rdma_cm.h
+++ b/include/rdma/rdma_cm.h
@@ -111,6 +111,20 @@ struct rdma_cm_event {
         } param;
 };
 
+enum rdma_cm_state {
+        RDMA_CM_IDLE,
+        RDMA_CM_ADDR_QUERY,
+        RDMA_CM_ADDR_RESOLVED,
+        RDMA_CM_ROUTE_QUERY,
+        RDMA_CM_ROUTE_RESOLVED,
+        RDMA_CM_CONNECT,
+        RDMA_CM_DISCONNECT,
+        RDMA_CM_ADDR_BOUND,
+        RDMA_CM_LISTEN,
+        RDMA_CM_DEVICE_REMOVAL,
+        RDMA_CM_DESTROYING
+};
+
 struct rdma_cm_id;
 
 /**
@@ -130,6 +144,7 @@ struct rdma_cm_id {
         rdma_cm_event_handler    event_handler;
         struct rdma_route        route;
         enum rdma_port_space     ps;
+        enum ib_qp_type          qp_type;
         u8                       port_num;
 };
 
@@ -140,9 +155,11 @@ struct rdma_cm_id {
  *   returned rdma_id.
  * @context: User specified context associated with the id.
  * @ps: RDMA port space.
+ * @qp_type: type of queue pair associated with the id.
  */
 struct rdma_cm_id *rdma_create_id(rdma_cm_event_handler event_handler,
-                                  void *context, enum rdma_port_space ps);
+                                  void *context, enum rdma_port_space ps,
+                                  enum ib_qp_type qp_type);
 
 /**
   * rdma_destroy_id - Destroys an RDMA identifier.
@@ -329,4 +346,14 @@ void rdma_leave_multicast(struct rdma_cm_id *id, struct sockaddr *addr);
  */
 void rdma_set_service_type(struct rdma_cm_id *id, int tos);
 
+/**
+ * rdma_set_reuseaddr - Allow the reuse of local addresses when binding
+ *    the rdma_cm_id.
+ * @id: Communication identifier to configure.
+ * @reuse: Value indicating if the bound address is reusable.
+ *
+ * Reuse must be set before an address is bound to the id.
+ */
+int rdma_set_reuseaddr(struct rdma_cm_id *id, int reuse);
+
 #endif /* RDMA_CM_H */
diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h
new file mode 100644
index 000000000000..3c5363ab867b
--- /dev/null
+++ b/include/rdma/rdma_netlink.h
@@ -0,0 +1,92 @@
+#ifndef _RDMA_NETLINK_H
+#define _RDMA_NETLINK_H
+
+#include <linux/types.h>
+
+enum {
+        RDMA_NL_RDMA_CM = 1
+};
+
+#define RDMA_NL_GET_CLIENT(type) ((type & (((1 << 6) - 1) << 10)) >> 10)
+#define RDMA_NL_GET_OP(type) (type & ((1 << 10) - 1))
+#define RDMA_NL_GET_TYPE(client, op) ((client << 10) + op)
+
+enum {
+        RDMA_NL_RDMA_CM_ID_STATS = 0,
+        RDMA_NL_RDMA_CM_NUM_OPS
+};
+
+enum {
+        RDMA_NL_RDMA_CM_ATTR_SRC_ADDR = 1,
+        RDMA_NL_RDMA_CM_ATTR_DST_ADDR,
+        RDMA_NL_RDMA_CM_NUM_ATTR,
+};
+
+struct rdma_cm_id_stats {
+        __u32   qp_num;
+        __u32   bound_dev_if;
+        __u32   port_space;
+        __s32   pid;
+        __u8    cm_state;
+        __u8    node_type;
+        __u8    port_num;
+        __u8    qp_type;
+};
+
+#ifdef __KERNEL__
+
+#include <linux/netlink.h>
+
+struct ibnl_client_cbs {
+        int (*dump)(struct sk_buff *skb, struct netlink_callback *nlcb);
+};
+
+int ibnl_init(void);
+void ibnl_cleanup(void);
+
+/**
+ * Add a a client to the list of IB netlink exporters.
+ * @index: Index of the added client
+ * @nops: Number of supported ops by the added client.
+ * @cb_table: A table for op->callback
+ *
+ * Returns 0 on success or a negative error code.
+ */
+int ibnl_add_client(int index, int nops,
+                    const struct ibnl_client_cbs cb_table[]);
+
+/**
+ * Remove a client from IB netlink.
+ * @index: Index of the removed IB client.
+ *
+ * Returns 0 on success or a negative error code.
+ */
+int ibnl_remove_client(int index);
+
+/**
+ * Put a new message in a supplied skb.
+ * @skb: The netlink skb.
+ * @nlh: Pointer to put the header of the new netlink message.
+ * @seq: The message sequence number.
+ * @len: The requested message length to allocate.
+ * @client: Calling IB netlink client.
+ * @op: message content op.
+ * Returns the allocated buffer on success and NULL on failure.
+ */
+void *ibnl_put_msg(struct sk_buff *skb, struct nlmsghdr **nlh, int seq,
+                   int len, int client, int op);
+/**
+ * Put a new attribute in a supplied skb.
+ * @skb: The netlink skb.
+ * @nlh: Header of the netlink message to append the attribute to.
+ * @len: The length of the attribute data.
+ * @data: The attribute data to put.
+ * @type: The attribute type.
+ * Returns the 0 and a negative error code on failure.
+ */
+int ibnl_put_attr(struct sk_buff *skb, struct nlmsghdr *nlh,
+                  int len, void *data, int type);
+
+#endif /* __KERNEL__ */
+
+#endif /* _RDMA_NETLINK_H */
diff --git a/include/rdma/rdma_user_cm.h b/include/rdma/rdma_user_cm.h
index 1d165022c02d..fc82c1896f75 100644
--- a/include/rdma/rdma_user_cm.h
+++ b/include/rdma/rdma_user_cm.h
@@ -221,8 +221,9 @@ enum {
 
 /* Option details */
 enum {
-        RDMA_OPTION_ID_TOS      = 0,
-        RDMA_OPTION_IB_PATH     = 1
+        RDMA_OPTION_ID_TOS       = 0,
+        RDMA_OPTION_ID_REUSEADDR = 1,
+        RDMA_OPTION_IB_PATH      = 1
 };
 
 struct rdma_ucm_set_option {