Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  RDMA/cma: Save PID of ID's owner
  RDMA/cma: Add support for netlink statistics export
  RDMA/cma: Pass QP type into rdma_create_id()
  RDMA: Update exported headers list
  RDMA/cma: Export enum cma_state in <rdma/rdma_cm.h>
  RDMA/nes: Add a check for strict_strtoul()
  RDMA/cxgb3: Don't post zero-byte read if endpoint is going away
  RDMA/cxgb4: Use completion objects for event blocking
  IB/srp: Fix integer -> pointer cast warnings
  IB: Add devnode methods to cm_class and umad_class
  IB/mad: Return EPROTONOSUPPORT when an RDMA device lacks the QP required
  IB/uverbs: Add devnode method to set path/mode
  RDMA/ucma: Add .nodename/.mode to tell userspace where to create device node
  RDMA: Add netlink infrastructure
  RDMA: Add error handling to ib_core_init()

drivers/infiniband/Kconfig

@@ -2,6 +2,7 @@ menuconfig INFINIBAND
 	tristate "InfiniBand support"
 	depends on PCI || BROKEN
 	depends on HAS_IOMEM
+	depends on NET
 	---help---
 	  Core support for InfiniBand (IB).  Make sure to also select
 	  any protocols you wish to use as well as drivers for your

drivers/infiniband/core/Makefile

@@ -8,7 +8,7 @@ obj-$(CONFIG_INFINIBAND_USER_ACCESS) +=	ib_uverbs.o ib_ucm.o \
 					$(user_access-y)
 
 ib_core-y :=			packer.o ud_header.o verbs.o sysfs.o \
-				device.o fmr_pool.o cache.o
+				device.o fmr_pool.o cache.o netlink.o
 ib_core-$(CONFIG_INFINIBAND_USER_MEM) += umem.o
 
 ib_mad-y :=			mad.o smi.o agent.o mad_rmpp.o

drivers/infiniband/core/cm.c

@@ -3639,8 +3639,16 @@ static struct kobj_type cm_port_obj_type = {
 	.release = cm_release_port_obj
 };
 
+static char *cm_devnode(struct device *dev, mode_t *mode)
+{
+	*mode = 0666;
+	return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev));
+}
+
 struct class cm_class = {
+	.owner   = THIS_MODULE,
 	.name    = "infiniband_cm",
+	.devnode = cm_devnode,
 };
 EXPORT_SYMBOL(cm_class);
 

drivers/infiniband/core/device.c

@@ -38,6 +38,7 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/mutex.h>
+#include <rdma/rdma_netlink.h>
 
 #include "core_priv.h"
 
@@ -725,22 +726,40 @@ static int __init ib_core_init(void)
 		return -ENOMEM;
 
 	ret = ib_sysfs_setup();
-	if (ret)
+	if (ret) {
 		printk(KERN_WARNING "Couldn't create InfiniBand device class\n");
+		goto err;
+	}
+
+	ret = ibnl_init();
+	if (ret) {
+		printk(KERN_WARNING "Couldn't init IB netlink interface\n");
+		goto err_sysfs;
+	}
 
 	ret = ib_cache_setup();
 	if (ret) {
 		printk(KERN_WARNING "Couldn't set up InfiniBand P_Key/GID cache\n");
-		ib_sysfs_cleanup();
-		destroy_workqueue(ib_wq);
+		goto err_nl;
 	}
 
+	return 0;
+
+err_nl:
+	ibnl_cleanup();
+
+err_sysfs:
+	ib_sysfs_cleanup();
+
+err:
+	destroy_workqueue(ib_wq);
 	return ret;
 }
 
 static void __exit ib_core_cleanup(void)
 {
 	ib_cache_cleanup();
+	ibnl_cleanup();
 	ib_sysfs_cleanup();
 	/* Make sure that any pending umem accounting work is done. */
 	destroy_workqueue(ib_wq);

drivers/infiniband/core/mad.c

@@ -276,6 +276,13 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
 		goto error1;
 	}
 
+	/* Verify the QP requested is supported.  For example, Ethernet devices
+	 * will not have QP0 */
+	if (!port_priv->qp_info[qpn].qp) {
+		ret = ERR_PTR(-EPROTONOSUPPORT);
+		goto error1;
+	}
+
 	/* Allocate structures */
 	mad_agent_priv = kzalloc(sizeof *mad_agent_priv, GFP_KERNEL);
 	if (!mad_agent_priv) {

drivers/infiniband/core/netlink.c

+/*
+ * Copyright (c) 2010 Voltaire Inc.  All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define pr_fmt(fmt) "%s:%s: " fmt, KBUILD_MODNAME, __func__
+
+#include <net/netlink.h>
+#include <net/net_namespace.h>
+#include <net/sock.h>
+#include <rdma/rdma_netlink.h>
+
+struct ibnl_client {
+	struct list_head		list;
+	int				index;
+	int				nops;
+	const struct ibnl_client_cbs   *cb_table;
+};
+
+static DEFINE_MUTEX(ibnl_mutex);
+static struct sock *nls;
+static LIST_HEAD(client_list);
+
+int ibnl_add_client(int index, int nops,
+		    const struct ibnl_client_cbs cb_table[])
+{
+	struct ibnl_client *cur;
+	struct ibnl_client *nl_client;
+
+	nl_client = kmalloc(sizeof *nl_client, GFP_KERNEL);
+	if (!nl_client)
+		return -ENOMEM;
+
+	nl_client->index	= index;
+	nl_client->nops		= nops;
+	nl_client->cb_table	= cb_table;
+
+	mutex_lock(&ibnl_mutex);
+
+	list_for_each_entry(cur, &client_list, list) {
+		if (cur->index == index) {
+			pr_warn("Client for %d already exists\n", index);
+			mutex_unlock(&ibnl_mutex);
+			kfree(nl_client);
+			return -EINVAL;
+		}
+	}
+
+	list_add_tail(&nl_client->list, &client_list);
+
+	mutex_unlock(&ibnl_mutex);
+
+	return 0;
+}
+EXPORT_SYMBOL(ibnl_add_client);
+
+int ibnl_remove_client(int index)
+{
+	struct ibnl_client *cur, *next;
+
+	mutex_lock(&ibnl_mutex);
+	list_for_each_entry_safe(cur, next, &client_list, list) {
+		if (cur->index == index) {
+			list_del(&(cur->list));
+			mutex_unlock(&ibnl_mutex);
+			kfree(cur);
+			return 0;
+		}
+	}
+	pr_warn("Can't remove callback for client idx %d. Not found\n", index);
+	mutex_unlock(&ibnl_mutex);
+
+	return -EINVAL;
+}
+EXPORT_SYMBOL(ibnl_remove_client);
+
+void *ibnl_put_msg(struct sk_buff *skb, struct nlmsghdr **nlh, int seq,
+		   int len, int client, int op)
+{
+	unsigned char *prev_tail;
+
+	prev_tail = skb_tail_pointer(skb);
+	*nlh = NLMSG_NEW(skb, 0, seq, RDMA_NL_GET_TYPE(client, op),
+			len, NLM_F_MULTI);
+	(*nlh)->nlmsg_len = skb_tail_pointer(skb) - prev_tail;
+	return NLMSG_DATA(*nlh);
+
+nlmsg_failure:
+	nlmsg_trim(skb, prev_tail);
+	return NULL;
+}
+EXPORT_SYMBOL(ibnl_put_msg);
+
+int ibnl_put_attr(struct sk_buff *skb, struct nlmsghdr *nlh,
+		  int len, void *data, int type)
+{
+	unsigned char *prev_tail;
+
+	prev_tail = skb_tail_pointer(skb);
+	NLA_PUT(skb, type, len, data);
+	nlh->nlmsg_len += skb_tail_pointer(skb) - prev_tail;
+	return 0;
+
+nla_put_failure:
+	nlmsg_trim(skb, prev_tail - nlh->nlmsg_len);
+	return -EMSGSIZE;
+}
+EXPORT_SYMBOL(ibnl_put_attr);
+
+static int ibnl_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh)
+{
+	struct ibnl_client *client;
+	int type = nlh->nlmsg_type;
+	int index = RDMA_NL_GET_CLIENT(type);
+	int op = RDMA_NL_GET_OP(type);
+
+	list_for_each_entry(client, &client_list, list) {
+		if (client->index == index) {
+			if (op < 0 || op >= client->nops ||
+			    !client->cb_table[RDMA_NL_GET_OP(op)].dump)
+				return -EINVAL;
+			return netlink_dump_start(nls, skb, nlh,
+						  client->cb_table[op].dump,
+						  NULL);
+		}
+	}
+
+	pr_info("Index %d wasn't found in client list\n", index);
+	return -EINVAL;
+}
+
+static void ibnl_rcv(struct sk_buff *skb)
+{
+	mutex_lock(&ibnl_mutex);
+	netlink_rcv_skb(skb, &ibnl_rcv_msg);
+	mutex_unlock(&ibnl_mutex);
+}
+
+int __init ibnl_init(void)
+{
+	nls = netlink_kernel_create(&init_net, NETLINK_RDMA, 0, ibnl_rcv,
+				    NULL, THIS_MODULE);
+	if (!nls) {
+		pr_warn("Failed to create netlink socket\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void ibnl_cleanup(void)
+{
+	struct ibnl_client *cur, *next;
+
+	mutex_lock(&ibnl_mutex);
+	list_for_each_entry_safe(cur, next, &client_list, list) {
+		list_del(&(cur->list));
+		kfree(cur);
+	}
+	mutex_unlock(&ibnl_mutex);
+
+	netlink_kernel_release(nls);
+}

drivers/infiniband/core/ucma.c

@@ -367,13 +367,28 @@ static ssize_t ucma_get_event(struct ucma_file *file, const char __user *inbuf,
 	return ret;
 }
 
-static ssize_t ucma_create_id(struct ucma_file *file,
-				const char __user *inbuf,
+static int ucma_get_qp_type(struct rdma_ucm_create_id *cmd, enum ib_qp_type *qp_type)
+{
+	switch (cmd->ps) {
+	case RDMA_PS_TCP:
+		*qp_type = IB_QPT_RC;
+		return 0;
+	case RDMA_PS_UDP:
+	case RDMA_PS_IPOIB:
+		*qp_type = IB_QPT_UD;
+		return 0;
+	default:
+		return -EINVAL;
+	}
+}
+
+static ssize_t ucma_create_id(struct ucma_file *file, const char __user *inbuf,
 			      int in_len, int out_len)
 {
 	struct rdma_ucm_create_id cmd;
 	struct rdma_ucm_create_id_resp resp;
 	struct ucma_context *ctx;
+	enum ib_qp_type qp_type;
 	int ret;
 
 	if (out_len < sizeof(resp))
@@ -382,6 +397,10 @@ static ssize_t ucma_create_id(struct ucma_file *file,
 	if (copy_from_user(&cmd, inbuf, sizeof(cmd)))
 		return -EFAULT;
 
+	ret = ucma_get_qp_type(&cmd, &qp_type);
+	if (ret)
+		return ret;
+
 	mutex_lock(&file->mut);
 	ctx = ucma_alloc_ctx(file);
 	mutex_unlock(&file->mut);
@@ -389,7 +408,7 @@ static ssize_t ucma_create_id(struct ucma_file *file,
 		return -ENOMEM;
 
 	ctx->uid = cmd.uid;
-	ctx->cm_id = rdma_create_id(ucma_event_handler, ctx, cmd.ps);
+	ctx->cm_id = rdma_create_id(ucma_event_handler, ctx, cmd.ps, qp_type);
 	if (IS_ERR(ctx->cm_id)) {
 		ret = PTR_ERR(ctx->cm_id);
 		goto err1;
@@ -1340,6 +1359,8 @@ static const struct file_operations ucma_fops = {
 static struct miscdevice ucma_misc = {
 	.minor		= MISC_DYNAMIC_MINOR,
 	.name		= "rdma_cm",
+	.nodename	= "infiniband/rdma_cm",
+	.mode		= 0666,
 	.fops		= &ucma_fops,
 };
 

drivers/infiniband/core/user_mad.c

@@ -1176,6 +1176,11 @@ static void ib_umad_remove_one(struct ib_device *device)
 	kref_put(&umad_dev->ref, ib_umad_release_dev);
 }
 
+static char *umad_devnode(struct device *dev, mode_t *mode)
+{
+	return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev));
+}
+
 static int __init ib_umad_init(void)
 {
 	int ret;
@@ -1194,6 +1199,8 @@ static int __init ib_umad_init(void)
 		goto out_chrdev;
 	}
 
+	umad_class->devnode = umad_devnode;
+
 	ret = class_create_file(umad_class, &class_attr_abi_version.attr);
 	if (ret) {
 		printk(KERN_ERR "user_mad: couldn't create abi_version attribute\n");

drivers/infiniband/core/uverbs_main.c

@@ -824,6 +824,12 @@ static void ib_uverbs_remove_one(struct ib_device *device)
 	kfree(uverbs_dev);
 }
 
+static char *uverbs_devnode(struct device *dev, mode_t *mode)
+{
+	*mode = 0666;
+	return kasprintf(GFP_KERNEL, "infiniband/%s", dev_name(dev));
+}
+
 static int __init ib_uverbs_init(void)
 {
 	int ret;
@@ -842,6 +848,8 @@ static int __init ib_uverbs_init(void)
 		goto out_chrdev;
 	}
 
+	uverbs_class->devnode = uverbs_devnode;
+
 	ret = class_create_file(uverbs_class, &class_attr_abi_version.attr);
 	if (ret) {
 		printk(KERN_ERR "user_verbs: couldn't create abi_version attribute\n");

drivers/infiniband/hw/cxgb3/iwch_cm.c

@@ -914,7 +914,7 @@ static void process_mpa_reply(struct iwch_ep *ep, struct sk_buff *skb)
 		goto err;
 
 	if (peer2peer && iwch_rqes_posted(ep->com.qp) == 0) {
-		iwch_post_zb_read(ep->com.qp);
+		iwch_post_zb_read(ep);
 	}
 
 	goto out;
@@ -1078,6 +1078,8 @@ static int tx_ack(struct t3cdev *tdev, struct sk_buff *skb, void *ctx)
 	struct iwch_ep *ep = ctx;
 	struct cpl_wr_ack *hdr = cplhdr(skb);
 	unsigned int credits = ntohs(hdr->credits);
+	unsigned long flags;
+	int post_zb = 0;
 
 	PDBG("%s ep %p credits %u\n", __func__, ep, credits);
 
@@ -1087,28 +1089,34 @@ static int tx_ack(struct t3cdev *tdev, struct sk_buff *skb, void *ctx)
 		return CPL_RET_BUF_DONE;
 	}
 
+	spin_lock_irqsave(&ep->com.lock, flags);
 	BUG_ON(credits != 1);
 	dst_confirm(ep->dst);
 	if (!ep->mpa_skb) {
 		PDBG("%s rdma_init wr_ack ep %p state %u\n",
-			__func__, ep, state_read(&ep->com));
+			__func__, ep, ep->com.state);
 		if (ep->mpa_attr.initiator) {
 			PDBG("%s initiator ep %p state %u\n",
-				__func__, ep, state_read(&ep->com));
-			if (peer2peer)
-				iwch_post_zb_read(ep->com.qp);
+				__func__, ep, ep->com.state);
+			if (peer2peer && ep->com.state == FPDU_MODE)
+				post_zb = 1;
 		} else {
 			PDBG("%s responder ep %p state %u\n",
-				__func__, ep, state_read(&ep->com));
+				__func__, ep, ep->com.state);
+			if (ep->com.state == MPA_REQ_RCVD) {
 				ep->com.rpl_done = 1;
 				wake_up(&ep->com.waitq);
 			}
+		}
 	} else {
 		PDBG("%s lsm ack ep %p state %u freeing skb\n",
-			__func__, ep, state_read(&ep->com));
+			__func__, ep, ep->com.state);
 		kfree_skb(ep->mpa_skb);
 		ep->mpa_skb = NULL;
 	}
+	spin_unlock_irqrestore(&ep->com.lock, flags);
+	if (post_zb)
+		iwch_post_zb_read(ep);
 	return CPL_RET_BUF_DONE;
 }
 

drivers/infiniband/hw/cxgb3/iwch_provider.h

@@ -332,7 +332,7 @@ int iwch_bind_mw(struct ib_qp *qp,
 			     struct ib_mw_bind *mw_bind);
 int iwch_poll_cq(struct ib_cq *ibcq, int num_entries, struct ib_wc *wc);
 int iwch_post_terminate(struct iwch_qp *qhp, struct respQ_msg_t *rsp_msg);
-int iwch_post_zb_read(struct iwch_qp *qhp);
+int iwch_post_zb_read(struct iwch_ep *ep);
 int iwch_register_device(struct iwch_dev *dev);
 void iwch_unregister_device(struct iwch_dev *dev);
 void stop_read_rep_timer(struct iwch_qp *qhp);

drivers/infiniband/hw/cxgb3/iwch_qp.c

@@ -738,7 +738,7 @@ static inline void build_term_codes(struct respQ_msg_t *rsp_msg,
 	}
 }
 
-int iwch_post_zb_read(struct iwch_qp *qhp)
+int iwch_post_zb_read(struct iwch_ep *ep)
 {
 	union t3_wr *wqe;
 	struct sk_buff *skb;
@@ -761,10 +761,10 @@ int iwch_post_zb_read(struct iwch_qp *qhp)
 	wqe->read.local_len = cpu_to_be32(0);
 	wqe->read.local_to = cpu_to_be64(1);
 	wqe->send.wrh.op_seop_flags = cpu_to_be32(V_FW_RIWR_OP(T3_WR_READ));
-	wqe->send.wrh.gen_tid_len = cpu_to_be32(V_FW_RIWR_TID(qhp->ep->hwtid)|
+	wqe->send.wrh.gen_tid_len = cpu_to_be32(V_FW_RIWR_TID(ep->hwtid)|
 						V_FW_RIWR_LEN(flit_cnt));
 	skb->priority = CPL_PRIORITY_DATA;
-	return iwch_cxgb3_ofld_send(qhp->rhp->rdev.t3cdev_p, skb);
+	return iwch_cxgb3_ofld_send(ep->com.qp->rhp->rdev.t3cdev_p, skb);
 }
 
 /*

drivers/infiniband/hw/cxgb4/iw_cxgb4.h

@@ -35,7 +35,7 @@
 #include <linux/list.h>
 #include <linux/spinlock.h>
 #include <linux/idr.h>
-#include <linux/workqueue.h>
+#include <linux/completion.h>
 #include <linux/netdevice.h>
 #include <linux/sched.h>
 #include <linux/pci.h>
@@ -131,28 +131,21 @@ static inline int c4iw_num_stags(struct c4iw_rdev *rdev)
 
 #define C4IW_WR_TO (10*HZ)
 
-enum {
-	REPLY_READY = 0,
-};
-
 struct c4iw_wr_wait {
-	wait_queue_head_t wait;
-	unsigned long status;
+	struct completion completion;
 	int ret;
 };
 
 static inline void c4iw_init_wr_wait(struct c4iw_wr_wait *wr_waitp)
 {
 	wr_waitp->ret = 0;
-	wr_waitp->status = 0;
-	init_waitqueue_head(&wr_waitp->wait);
+	init_completion(&wr_waitp->completion);
 }
 
 static inline void c4iw_wake_up(struct c4iw_wr_wait *wr_waitp, int ret)
 {
 	wr_waitp->ret = ret;
-	set_bit(REPLY_READY, &wr_waitp->status);
-	wake_up(&wr_waitp->wait);
+	complete(&wr_waitp->completion);
 }
 
 static inline int c4iw_wait_for_reply(struct c4iw_rdev *rdev,
@@ -164,8 +157,7 @@ static inline int c4iw_wait_for_reply(struct c4iw_rdev *rdev,
 	int ret;
 
 	do {
-		ret = wait_event_timeout(wr_waitp->wait,
-			test_and_clear_bit(REPLY_READY, &wr_waitp->status), to);
+		ret = wait_for_completion_timeout(&wr_waitp->completion, to);
 		if (!ret) {
 			printk(KERN_ERR MOD "%s - Device %s not responding - "
 			       "tid %u qpid %u\n", func,

drivers/infiniband/hw/nes/nes.c

@@ -1138,7 +1138,9 @@ static ssize_t nes_store_wqm_quanta(struct device_driver *ddp,
 	u32 i = 0;
 	struct nes_device *nesdev;
 
-	strict_strtoul(buf, 0, &wqm_quanta_value);
+	if (kstrtoul(buf, 0, &wqm_quanta_value) < 0)
+		return -EINVAL;
+
 	list_for_each_entry(nesdev, &nes_dev_list, list) {
 		if (i == ee_flsh_adapter) {
 			nesdev->nesadapter->wqm_quanta = wqm_quanta_value;

drivers/infiniband/hw/qib/Kconfig

 config INFINIBAND_QIB
 	tristate "QLogic PCIe HCA support"
-	depends on 64BIT && NET
+	depends on 64BIT
 	---help---
 	This is a low-level driver for QLogic PCIe QLE InfiniBand host
 	channel adapters.  This driver does not support the QLogic

drivers/infiniband/ulp/iser/iser_verbs.c

@@ -548,7 +548,7 @@ int iser_connect(struct iser_conn   *ib_conn,
 	iser_conn_get(ib_conn); /* ref ib conn's cma id */
 	ib_conn->cma_id = rdma_create_id(iser_cma_handler,
 					     (void *)ib_conn,
-					     RDMA_PS_TCP);
+					     RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(ib_conn->cma_id)) {
 		err = PTR_ERR(ib_conn->cma_id);
 		iser_err("rdma_create_id failed: %d\n", err);

drivers/infiniband/ulp/srp/ib_srp.c

@@ -1147,7 +1147,7 @@ static void srp_process_aer_req(struct srp_target_port *target,
 static void srp_handle_recv(struct srp_target_port *target, struct ib_wc *wc)
 {
 	struct ib_device *dev = target->srp_host->srp_dev->dev;
-	struct srp_iu *iu = (struct srp_iu *) wc->wr_id;
+	struct srp_iu *iu = (struct srp_iu *) (uintptr_t) wc->wr_id;
 	int res;
 	u8 opcode;
 
@@ -1231,7 +1231,7 @@ static void srp_send_completion(struct ib_cq *cq, void *target_ptr)
 			break;
 		}
 
-		iu = (struct srp_iu *) wc.wr_id;
+		iu = (struct srp_iu *) (uintptr_t) wc.wr_id;
 		list_add(&iu->list, &target->free_tx);
 	}
 }

include/linux/netlink.h

@@ -24,6 +24,7 @@
 /* leave room for NETLINK_DM (DM Events) */
 #define NETLINK_SCSITRANSPORT	18	/* SCSI Transports */
 #define NETLINK_ECRYPTFS	19
+#define NETLINK_RDMA		20
 
 #define MAX_LINKS 32		
 

include/rdma/Kbuild

+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

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

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.

include/rdma/rdma_netlink.h

+#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 */

net/9p/trans_rdma.c

@@ -589,7 +589,8 @@ rdma_create_trans(struct p9_client *client, const char *addr, char *args)
 		return -ENOMEM;
 
 	/* Create the RDMA CM ID */
-	rdma->cm_id = rdma_create_id(p9_cm_event_handler, client, RDMA_PS_TCP);
+	rdma->cm_id = rdma_create_id(p9_cm_event_handler, client, RDMA_PS_TCP,
+				     IB_QPT_RC);
 	if (IS_ERR(rdma->cm_id))
 		goto error;
 

net/rds/ib.c

@@ -325,7 +325,7 @@ static int rds_ib_laddr_check(__be32 addr)
 	/* Create a CMA ID and try to bind it. This catches both
 	 * IB and iWARP capable NICs.
 	 */
-	cm_id = rdma_create_id(NULL, NULL, RDMA_PS_TCP);
+	cm_id = rdma_create_id(NULL, NULL, RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(cm_id))
 		return PTR_ERR(cm_id);
 

net/rds/ib_cm.c

@@ -587,7 +587,7 @@ int rds_ib_conn_connect(struct rds_connection *conn)
 	/* XXX I wonder what affect the port space has */
 	/* delegate cm event handler to rdma_transport */
 	ic->i_cm_id = rdma_create_id(rds_rdma_cm_event_handler, conn,
-				     RDMA_PS_TCP);
+				     RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(ic->i_cm_id)) {
 		ret = PTR_ERR(ic->i_cm_id);
 		ic->i_cm_id = NULL;

net/rds/iw.c

@@ -226,7 +226,7 @@ static int rds_iw_laddr_check(__be32 addr)
 	/* Create a CMA ID and try to bind it. This catches both
 	 * IB and iWARP capable NICs.
 	 */
-	cm_id = rdma_create_id(NULL, NULL, RDMA_PS_TCP);
+	cm_id = rdma_create_id(NULL, NULL, RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(cm_id))
 		return PTR_ERR(cm_id);
 

net/rds/iw_cm.c

@@ -522,7 +522,7 @@ int rds_iw_conn_connect(struct rds_connection *conn)
 	/* XXX I wonder what affect the port space has */
 	/* delegate cm event handler to rdma_transport */
 	ic->i_cm_id = rdma_create_id(rds_rdma_cm_event_handler, conn,
-				     RDMA_PS_TCP);
+				     RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(ic->i_cm_id)) {
 		ret = PTR_ERR(ic->i_cm_id);
 		ic->i_cm_id = NULL;

net/rds/rdma_transport.c

@@ -158,7 +158,8 @@ static int rds_rdma_listen_init(void)
 	struct rdma_cm_id *cm_id;
 	int ret;
 
-	cm_id = rdma_create_id(rds_rdma_cm_event_handler, NULL, RDMA_PS_TCP);
+	cm_id = rdma_create_id(rds_rdma_cm_event_handler, NULL, RDMA_PS_TCP,
+			       IB_QPT_RC);
 	if (IS_ERR(cm_id)) {
 		ret = PTR_ERR(cm_id);
 		printk(KERN_ERR "RDS/RDMA: failed to setup listener, "

net/sunrpc/xprtrdma/svc_rdma_transport.c

@@ -695,7 +695,8 @@ static struct svc_xprt *svc_rdma_create(struct svc_serv *serv,
 		return ERR_PTR(-ENOMEM);
 	xprt = &cma_xprt->sc_xprt;
 
-	listen_id = rdma_create_id(rdma_listen_handler, cma_xprt, RDMA_PS_TCP);
+	listen_id = rdma_create_id(rdma_listen_handler, cma_xprt, RDMA_PS_TCP,
+				   IB_QPT_RC);
 	if (IS_ERR(listen_id)) {
 		ret = PTR_ERR(listen_id);
 		dprintk("svcrdma: rdma_create_id failed = %d\n", ret);

net/sunrpc/xprtrdma/verbs.c

@@ -387,7 +387,7 @@ rpcrdma_create_id(struct rpcrdma_xprt *xprt,
 
 	init_completion(&ia->ri_done);
 
-	id = rdma_create_id(rpcrdma_conn_upcall, xprt, RDMA_PS_TCP);
+	id = rdma_create_id(rpcrdma_conn_upcall, xprt, RDMA_PS_TCP, IB_QPT_RC);
 	if (IS_ERR(id)) {
 		rc = PTR_ERR(id);
 		dprintk("RPC:       %s: rdma_create_id() failed %i\n",