RDMA subsystem updates for 5.4

This cycle mainly saw lots of bug fixes and clean up code across the core
 code and several drivers, few new functional changes were made.
 
 - Many cleanup and bug fixes for hns
 
 - Various small bug fixes and cleanups in hfi1, mlx5, usnic, qed,
   bnxt_re, efa
 
 - Share the query_port code between all the iWarp drivers
 
 - General rework and cleanup of the ODP MR umem code to fit better with
   the mmu notifier get/put scheme
 
 - Support rdma netlink in non init_net name spaces
 
 - mlx5 support for XRC devx and DC ODP
 -----BEGIN PGP SIGNATURE-----
 
 iQIzBAABCgAdFiEEfB7FMLh+8QxL+6i3OG33FX4gmxoFAl2A1ugACgkQOG33FX4g
 mxp+EQ//Ug8CyyDs40SGZztItoGghuQ4TVA2pXtuZ9LkFVJRWhYPJGOadHIYXqGO
 KQJJpZPQ02HTZUPWBZNKmD5bwHfErm4cS73/mVmUpximnUqu/UsLRJp8SIGmBg1D
 W1Lz1BJX24MdV8aUnChYvdL5Hbl52q+BaE99Z0haOvW7I3YnKJC34mR8m/A5MiRf
 rsNIZNPHdq2U6pKLgCbOyXib8yBcJQqBb8x4WNAoB1h4MOx+ir5QLfr3ozrZs1an
 xXgqyiOBmtkUgCMIpXC4juPN/6gw3Y5nkk2VIWY+MAY1a7jZPbI+6LJZZ1Uk8R44
 Lf2KSzabFMMYGKJYE1Znxk+JWV8iE+m+n6wWEfRM9l0b4gXXbrtKgaouFbsLcsQA
 CvBEQuiFSO9Kq01JPaAN1XDmhqyTarG6lHjXnW7ifNlLXnPbR1RJlprycExOhp0c
 axum5K2fRNW2+uZJt+zynMjk2kyjT1lnlsr1Rbgc4Pyionaiydg7zwpiac7y/bdS
 F7/WqdmPiff78KMi187EF5pjFqMWhthvBtTbWDuuxaxc2nrXSdiCUN+96j1amQUH
 yU/7AZzlnKeKEQQCR4xddsSs2eTrXiLLFRLk9GCK2eh4cUN212eHTrPLKkQf1cN+
 ydYbR2pxw3B38LCCNBy+bL+u7e/Tyehs4ynATMpBuEgc5iocTwE=
 =zHXW
 -----END PGP SIGNATURE-----

Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma

Pull RDMA subsystem updates from Jason Gunthorpe:
 "This cycle mainly saw lots of bug fixes and clean up code across the
  core code and several drivers, few new functional changes were made.

   - Many cleanup and bug fixes for hns

   - Various small bug fixes and cleanups in hfi1, mlx5, usnic, qed,
     bnxt_re, efa

   - Share the query_port code between all the iWarp drivers

   - General rework and cleanup of the ODP MR umem code to fit better
     with the mmu notifier get/put scheme

   - Support rdma netlink in non init_net name spaces

   - mlx5 support for XRC devx and DC ODP"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma: (99 commits)
  RDMA: Fix double-free in srq creation error flow
  RDMA/efa: Fix incorrect error print
  IB/mlx5: Free mpi in mp_slave mode
  IB/mlx5: Use the original address for the page during free_pages
  RDMA/bnxt_re: Fix spelling mistake "missin_resp" -> "missing_resp"
  RDMA/hns: Package operations of rq inline buffer into separate functions
  RDMA/hns: Optimize cmd init and mode selection for hip08
  IB/hfi1: Define variables as unsigned long to fix KASAN warning
  IB/{rdmavt, hfi1, qib}: Add a counter for credit waits
  IB/hfi1: Add traces for TID RDMA READ
  RDMA/siw: Relax from kmap_atomic() use in TX path
  IB/iser: Support up to 16MB data transfer in a single command
  RDMA/siw: Fix page address mapping in TX path
  RDMA: Fix goto target to release the allocated memory
  RDMA/usnic: Avoid overly large buffers on stack
  RDMA/odp: Add missing cast for 32 bit
  RDMA/hns: Use devm_platform_ioremap_resource() to simplify code
  Documentation/infiniband: update name of some functions
  RDMA/cma: Fix false error message
  RDMA/hns: Fix wrong assignment of qp_access_flags
  ...

Documentation/infiniband/core_locking.rst

@@ -29,10 +29,10 @@ Sleeping and interrupt context
   The corresponding functions exported to upper level protocol
   consumers:
 
-    - ib_create_ah
+    - rdma_create_ah
-    - ib_modify_ah
+    - rdma_modify_ah
-    - ib_query_ah
+    - rdma_query_ah
-    - ib_destroy_ah
+    - rdma_destroy_ah
     - ib_post_send
     - ib_post_recv
     - ib_req_notify_cq

drivers/infiniband/core/addr.c

@@ -183,7 +183,7 @@ static int ib_nl_ip_send_msg(struct rdma_dev_addr *dev_addr,
 
 	/* Repair the nlmsg header length */
 	nlmsg_end(skb, nlh);
-	rdma_nl_multicast(skb, RDMA_NL_GROUP_LS, GFP_KERNEL);
+	rdma_nl_multicast(&init_net, skb, RDMA_NL_GROUP_LS, GFP_KERNEL);
 
 	/* Make the request retry, so when we get the response from userspace
 	 * we will have something.

drivers/infiniband/core/cache.c

@@ -810,6 +810,7 @@ static void release_gid_table(struct ib_device *device,
 	if (leak)
 		return;
 
+	mutex_destroy(&table->lock);
 	kfree(table->data_vec);
 	kfree(table);
 }

drivers/infiniband/core/cma.c

@@ -3046,7 +3046,7 @@ static void addr_handler(int status, struct sockaddr *src_addr,
 		if (status)
 			pr_debug_ratelimited("RDMA CM: ADDR_ERROR: failed to acquire device. status %d\n",
 					     status);
-	} else {
+	} else if (status) {
 		pr_debug_ratelimited("RDMA CM: ADDR_ERROR: failed to resolve IP. status %d\n", status);
 	}
 

drivers/infiniband/core/cma_configfs.c

@@ -342,12 +342,18 @@ static struct configfs_subsystem cma_subsys = {
 
 int __init cma_configfs_init(void)
 {
+	int ret;
+
 	config_group_init(&cma_subsys.su_group);
 	mutex_init(&cma_subsys.su_mutex);
-	return configfs_register_subsystem(&cma_subsys);
+	ret = configfs_register_subsystem(&cma_subsys);
+	if (ret)
+		mutex_destroy(&cma_subsys.su_mutex);
+	return ret;
 }
 
 void __exit cma_configfs_exit(void)
 {
 	configfs_unregister_subsystem(&cma_subsys);
+	mutex_destroy(&cma_subsys.su_mutex);
 }

drivers/infiniband/core/core_priv.h

@@ -36,6 +36,8 @@
 #include <linux/list.h>
 #include <linux/spinlock.h>
 #include <linux/cgroup_rdma.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
 
 #include <rdma/ib_verbs.h>
 #include <rdma/opa_addr.h>
@@ -54,8 +56,26 @@ struct pkey_index_qp_list {
 	struct list_head    qp_list;
 };
 
+/**
+ * struct rdma_dev_net - rdma net namespace metadata for a net
+ * @nl_sock:	Pointer to netlink socket
+ * @net:	Pointer to owner net namespace
+ * @id:		xarray id to identify the net namespace.
+ */
+struct rdma_dev_net {
+	struct sock *nl_sock;
+	possible_net_t net;
+	u32 id;
+};
+
 extern const struct attribute_group ib_dev_attr_group;
 extern bool ib_devices_shared_netns;
+extern unsigned int rdma_dev_net_id;
+
+static inline struct rdma_dev_net *rdma_net_to_dev_net(struct net *net)
+{
+	return net_generic(net, rdma_dev_net_id);
+}
 
 int ib_device_register_sysfs(struct ib_device *device);
 void ib_device_unregister_sysfs(struct ib_device *device);
@@ -179,7 +199,6 @@ void ib_mad_cleanup(void);
 int ib_sa_init(void);
 void ib_sa_cleanup(void);
 
-int rdma_nl_init(void);
 void rdma_nl_exit(void);
 
 int ib_nl_handle_resolve_resp(struct sk_buff *skb,
@@ -365,4 +384,7 @@ void ib_port_unregister_module_stat(struct kobject *kobj);
 
 int ib_device_set_netns_put(struct sk_buff *skb,
 			    struct ib_device *dev, u32 ns_fd);
+
+int rdma_nl_net_init(struct rdma_dev_net *rnet);
+void rdma_nl_net_exit(struct rdma_dev_net *rnet);
 #endif /* _CORE_PRIV_H */

drivers/infiniband/core/counters.c

@@ -599,7 +599,7 @@ int rdma_counter_get_mode(struct ib_device *dev, u8 port,
 void rdma_counter_init(struct ib_device *dev)
 {
 	struct rdma_port_counter *port_counter;
-	u32 port;
+	u32 port, i;
 
 	if (!dev->port_data)
 		return;
@@ -620,13 +620,12 @@ void rdma_counter_init(struct ib_device *dev)
 	return;
 
 fail:
-	rdma_for_each_port(dev, port) {
+	for (i = port; i >= rdma_start_port(dev); i--) {
 		port_counter = &dev->port_data[port].port_counter;
 		kfree(port_counter->hstats);
 		port_counter->hstats = NULL;
+		mutex_destroy(&port_counter->lock);
 	}
-
-	return;
 }
 
 void rdma_counter_release(struct ib_device *dev)
@@ -637,5 +636,6 @@ void rdma_counter_release(struct ib_device *dev)
 	rdma_for_each_port(dev, port) {
 		port_counter = &dev->port_data[port].port_counter;
 		kfree(port_counter->hstats);
+		mutex_destroy(&port_counter->lock);
 	}
 }

drivers/infiniband/core/cq.c

@@ -252,6 +252,34 @@ out_free_cq:
 }
 EXPORT_SYMBOL(__ib_alloc_cq_user);
 
+/**
+ * __ib_alloc_cq_any - allocate a completion queue
+ * @dev:		device to allocate the CQ for
+ * @private:		driver private data, accessible from cq->cq_context
+ * @nr_cqe:		number of CQEs to allocate
+ * @poll_ctx:		context to poll the CQ from
+ * @caller:		module owner name
+ *
+ * Attempt to spread ULP Completion Queues over each device's interrupt
+ * vectors. A simple best-effort mechanism is used.
+ */
+struct ib_cq *__ib_alloc_cq_any(struct ib_device *dev, void *private,
+				int nr_cqe, enum ib_poll_context poll_ctx,
+				const char *caller)
+{
+	static atomic_t counter;
+	int comp_vector = 0;
+
+	if (dev->num_comp_vectors > 1)
+		comp_vector =
+			atomic_inc_return(&counter) %
+			min_t(int, dev->num_comp_vectors, num_online_cpus());
+
+	return __ib_alloc_cq_user(dev, private, nr_cqe, comp_vector, poll_ctx,
+				  caller, NULL);
+}
+EXPORT_SYMBOL(__ib_alloc_cq_any);
+
 /**
  * ib_free_cq_user - free a completion queue
  * @cq:		completion queue to free.

drivers/infiniband/core/device.c

@@ -39,7 +39,6 @@
 #include <linux/init.h>
 #include <linux/netdevice.h>
 #include <net/net_namespace.h>
-#include <net/netns/generic.h>
 #include <linux/security.h>
 #include <linux/notifier.h>
 #include <linux/hashtable.h>
@@ -111,17 +110,7 @@ static void ib_client_put(struct ib_client *client)
  */
 #define CLIENT_DATA_REGISTERED XA_MARK_1
 
-/**
+unsigned int rdma_dev_net_id;
- * struct rdma_dev_net - rdma net namespace metadata for a net
- * @net:	Pointer to owner net namespace
- * @id:		xarray id to identify the net namespace.
- */
-struct rdma_dev_net {
-	possible_net_t net;
-	u32 id;
-};
-
-static unsigned int rdma_dev_net_id;
 
 /*
  * A list of net namespaces is maintained in an xarray. This is necessary
@@ -514,6 +503,9 @@ static void ib_device_release(struct device *device)
 			  rcu_head);
 	}
 
+	mutex_destroy(&dev->unregistration_lock);
+	mutex_destroy(&dev->compat_devs_mutex);
+
 	xa_destroy(&dev->compat_devs);
 	xa_destroy(&dev->client_data);
 	kfree_rcu(dev, rcu_head);
@@ -1060,7 +1052,7 @@ int rdma_compatdev_set(u8 enable)
 
 static void rdma_dev_exit_net(struct net *net)
 {
-	struct rdma_dev_net *rnet = net_generic(net, rdma_dev_net_id);
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
 	struct ib_device *dev;
 	unsigned long index;
 	int ret;
@@ -1094,25 +1086,32 @@ static void rdma_dev_exit_net(struct net *net)
 	}
 	up_read(&devices_rwsem);
 
+	rdma_nl_net_exit(rnet);
 	xa_erase(&rdma_nets, rnet->id);
 }
 
 static __net_init int rdma_dev_init_net(struct net *net)
 {
-	struct rdma_dev_net *rnet = net_generic(net, rdma_dev_net_id);
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
 	unsigned long index;
 	struct ib_device *dev;
 	int ret;
 
+	write_pnet(&rnet->net, net);
+
+	ret = rdma_nl_net_init(rnet);
+	if (ret)
+		return ret;
+
 	/* No need to create any compat devices in default init_net. */
 	if (net_eq(net, &init_net))
 		return 0;
 
-	write_pnet(&rnet->net, net);
-
 	ret = xa_alloc(&rdma_nets, &rnet->id, rnet, xa_limit_32b, GFP_KERNEL);
-	if (ret)
+	if (ret) {
+		rdma_nl_net_exit(rnet);
 		return ret;
+	}
 
 	down_read(&devices_rwsem);
 	xa_for_each_marked (&devices, index, dev, DEVICE_REGISTERED) {
@@ -1974,6 +1973,74 @@ void ib_dispatch_event(struct ib_event *event)
 }
 EXPORT_SYMBOL(ib_dispatch_event);
 
+static int iw_query_port(struct ib_device *device,
+			   u8 port_num,
+			   struct ib_port_attr *port_attr)
+{
+	struct in_device *inetdev;
+	struct net_device *netdev;
+	int err;
+
+	memset(port_attr, 0, sizeof(*port_attr));
+
+	netdev = ib_device_get_netdev(device, port_num);
+	if (!netdev)
+		return -ENODEV;
+
+	dev_put(netdev);
+
+	port_attr->max_mtu = IB_MTU_4096;
+	port_attr->active_mtu = ib_mtu_int_to_enum(netdev->mtu);
+
+	if (!netif_carrier_ok(netdev)) {
+		port_attr->state = IB_PORT_DOWN;
+		port_attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;
+	} else {
+		inetdev = in_dev_get(netdev);
+
+		if (inetdev && inetdev->ifa_list) {
+			port_attr->state = IB_PORT_ACTIVE;
+			port_attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
+			in_dev_put(inetdev);
+		} else {
+			port_attr->state = IB_PORT_INIT;
+			port_attr->phys_state =
+				IB_PORT_PHYS_STATE_PORT_CONFIGURATION_TRAINING;
+		}
+	}
+
+	err = device->ops.query_port(device, port_num, port_attr);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+static int __ib_query_port(struct ib_device *device,
+			   u8 port_num,
+			   struct ib_port_attr *port_attr)
+{
+	union ib_gid gid = {};
+	int err;
+
+	memset(port_attr, 0, sizeof(*port_attr));
+
+	err = device->ops.query_port(device, port_num, port_attr);
+	if (err || port_attr->subnet_prefix)
+		return err;
+
+	if (rdma_port_get_link_layer(device, port_num) !=
+	    IB_LINK_LAYER_INFINIBAND)
+		return 0;
+
+	err = device->ops.query_gid(device, port_num, 0, &gid);
+	if (err)
+		return err;
+
+	port_attr->subnet_prefix = be64_to_cpu(gid.global.subnet_prefix);
+	return 0;
+}
+
 /**
  * ib_query_port - Query IB port attributes
  * @device:Device to query
@@ -1987,26 +2054,13 @@ int ib_query_port(struct ib_device *device,
 		  u8 port_num,
 		  struct ib_port_attr *port_attr)
 {
-	union ib_gid gid;
-	int err;
-
 	if (!rdma_is_port_valid(device, port_num))
 		return -EINVAL;
 
-	memset(port_attr, 0, sizeof(*port_attr));
+	if (rdma_protocol_iwarp(device, port_num))
-	err = device->ops.query_port(device, port_num, port_attr);
+		return iw_query_port(device, port_num, port_attr);
-	if (err || port_attr->subnet_prefix)
+	else
-		return err;
+		return __ib_query_port(device, port_num, port_attr);
-
-	if (rdma_port_get_link_layer(device, port_num) != IB_LINK_LAYER_INFINIBAND)
-		return 0;
-
-	err = device->ops.query_gid(device, port_num, 0, &gid);
-	if (err)
-		return err;
-
-	port_attr->subnet_prefix = be64_to_cpu(gid.global.subnet_prefix);
-	return 0;
 }
 EXPORT_SYMBOL(ib_query_port);
 
@@ -2661,12 +2715,6 @@ static int __init ib_core_init(void)
 		goto err_comp_unbound;
 	}
 
-	ret = rdma_nl_init();
-	if (ret) {
-		pr_warn("Couldn't init IB netlink interface: err %d\n", ret);
-		goto err_sysfs;
-	}
-
 	ret = addr_init();
 	if (ret) {
 		pr_warn("Could't init IB address resolution\n");
@@ -2712,8 +2760,6 @@ err_mad:
 err_addr:
 	addr_cleanup();
 err_ibnl:
-	rdma_nl_exit();
-err_sysfs:
 	class_unregister(&ib_class);
 err_comp_unbound:
 	destroy_workqueue(ib_comp_unbound_wq);

drivers/infiniband/core/fmr_pool.c

@@ -148,13 +148,6 @@ static void ib_fmr_batch_release(struct ib_fmr_pool *pool)
 		hlist_del_init(&fmr->cache_node);
 		fmr->remap_count = 0;
 		list_add_tail(&fmr->fmr->list, &fmr_list);
-
-#ifdef DEBUG
-		if (fmr->ref_count !=0) {
-			pr_warn(PFX "Unmapping FMR 0x%08x with ref count %d\n",
-				fmr, fmr->ref_count);
-		}
-#endif
 	}
 
 	list_splice_init(&pool->dirty_list, &unmap_list);
@@ -496,12 +489,6 @@ void ib_fmr_pool_unmap(struct ib_pool_fmr *fmr)
 		}
 	}
 
-#ifdef DEBUG
-	if (fmr->ref_count < 0)
-		pr_warn(PFX "FMR %p has ref count %d < 0\n",
-			fmr, fmr->ref_count);
-#endif
-
 	spin_unlock_irqrestore(&pool->pool_lock, flags);
 }
 EXPORT_SYMBOL(ib_fmr_pool_unmap);

drivers/infiniband/core/iwpm_msg.c

@@ -112,7 +112,7 @@ int iwpm_register_pid(struct iwpm_dev_data *pm_msg, u8 nl_client)
 	pr_debug("%s: Multicasting a nlmsg (dev = %s ifname = %s iwpm = %s)\n",
 		__func__, pm_msg->dev_name, pm_msg->if_name, iwpm_ulib_name);
 
-	ret = rdma_nl_multicast(skb, RDMA_NL_GROUP_IWPM, GFP_KERNEL);
+	ret = rdma_nl_multicast(&init_net, skb, RDMA_NL_GROUP_IWPM, GFP_KERNEL);
 	if (ret) {
 		skb = NULL; /* skb is freed in the netlink send-op handling */
 		iwpm_user_pid = IWPM_PID_UNAVAILABLE;
@@ -124,8 +124,7 @@ int iwpm_register_pid(struct iwpm_dev_data *pm_msg, u8 nl_client)
 	return ret;
 pid_query_error:
 	pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
-	if (skb)
+	dev_kfree_skb(skb);
-		dev_kfree_skb(skb);
 	if (nlmsg_request)
 		iwpm_free_nlmsg_request(&nlmsg_request->kref);
 	return ret;
@@ -202,7 +201,7 @@ int iwpm_add_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
 	nlmsg_end(skb, nlh);
 	nlmsg_request->req_buffer = pm_msg;
 
-	ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
+	ret = rdma_nl_unicast_wait(&init_net, skb, iwpm_user_pid);
 	if (ret) {
 		skb = NULL; /* skb is freed in the netlink send-op handling */
 		iwpm_user_pid = IWPM_PID_UNDEFINED;
@@ -214,8 +213,7 @@ int iwpm_add_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
 add_mapping_error:
 	pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
 add_mapping_error_nowarn:
-	if (skb)
+	dev_kfree_skb(skb);
-		dev_kfree_skb(skb);
 	if (nlmsg_request)
 		iwpm_free_nlmsg_request(&nlmsg_request->kref);
 	return ret;
@@ -297,7 +295,7 @@ int iwpm_add_and_query_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
 	nlmsg_end(skb, nlh);
 	nlmsg_request->req_buffer = pm_msg;
 
-	ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
+	ret = rdma_nl_unicast_wait(&init_net, skb, iwpm_user_pid);
 	if (ret) {
 		skb = NULL; /* skb is freed in the netlink send-op handling */
 		err_str = "Unable to send a nlmsg";
@@ -308,8 +306,7 @@ int iwpm_add_and_query_mapping(struct iwpm_sa_data *pm_msg, u8 nl_client)
 query_mapping_error:
 	pr_info("%s: %s (client = %d)\n", __func__, err_str, nl_client);
 query_mapping_error_nowarn:
-	if (skb)
+	dev_kfree_skb(skb);
-		dev_kfree_skb(skb);
 	if (nlmsg_request)
 		iwpm_free_nlmsg_request(&nlmsg_request->kref);
 	return ret;
@@ -364,7 +361,7 @@ int iwpm_remove_mapping(struct sockaddr_storage *local_addr, u8 nl_client)
 
 	nlmsg_end(skb, nlh);
 
-	ret = rdma_nl_unicast_wait(skb, iwpm_user_pid);
+	ret = rdma_nl_unicast_wait(&init_net, skb, iwpm_user_pid);
 	if (ret) {
 		skb = NULL; /* skb is freed in the netlink send-op handling */
 		iwpm_user_pid = IWPM_PID_UNDEFINED;

drivers/infiniband/core/iwpm_util.c

@@ -645,7 +645,7 @@ static int send_mapinfo_num(u32 mapping_num, u8 nl_client, int iwpm_pid)
 
 	nlmsg_end(skb, nlh);
 
-	ret = rdma_nl_unicast(skb, iwpm_pid);
+	ret = rdma_nl_unicast(&init_net, skb, iwpm_pid);
 	if (ret) {
 		skb = NULL;
 		err_str = "Unable to send a nlmsg";
@@ -655,8 +655,7 @@ static int send_mapinfo_num(u32 mapping_num, u8 nl_client, int iwpm_pid)
 	return 0;
 mapinfo_num_error:
 	pr_info("%s: %s\n", __func__, err_str);
-	if (skb)
+	dev_kfree_skb(skb);
-		dev_kfree_skb(skb);
 	return ret;
 }
 
@@ -674,7 +673,7 @@ static int send_nlmsg_done(struct sk_buff *skb, u8 nl_client, int iwpm_pid)
 		return -ENOMEM;
 	}
 	nlh->nlmsg_type = NLMSG_DONE;
-	ret = rdma_nl_unicast(skb, iwpm_pid);
+	ret = rdma_nl_unicast(&init_net, skb, iwpm_pid);
 	if (ret)
 		pr_warn("%s Unable to send a nlmsg\n", __func__);
 	return ret;
@@ -778,8 +777,7 @@ send_mapping_info_unlock:
 send_mapping_info_exit:
 	if (ret) {
 		pr_warn("%s: %s (ret = %d)\n", __func__, err_str, ret);
-		if (skb)
+		dev_kfree_skb(skb);
-			dev_kfree_skb(skb);
 		return ret;
 	}
 	send_nlmsg_done(skb, nl_client, iwpm_pid);
@@ -824,7 +822,7 @@ int iwpm_send_hello(u8 nl_client, int iwpm_pid, u16 abi_version)
 		goto hello_num_error;
 	nlmsg_end(skb, nlh);
 
-	ret = rdma_nl_unicast(skb, iwpm_pid);
+	ret = rdma_nl_unicast(&init_net, skb, iwpm_pid);
 	if (ret) {
 		skb = NULL;
 		err_str = "Unable to send a nlmsg";
@@ -834,7 +832,6 @@ int iwpm_send_hello(u8 nl_client, int iwpm_pid, u16 abi_version)
 	return 0;
 hello_num_error:
 	pr_info("%s: %s\n", __func__, err_str);
-	if (skb)
+	dev_kfree_skb(skb);
-		dev_kfree_skb(skb);
 	return ret;
 }

drivers/infiniband/core/netlink.c

@@ -36,20 +36,22 @@
 #include <linux/export.h>
 #include <net/netlink.h>
 #include <net/net_namespace.h>
+#include <net/netns/generic.h>
 #include <net/sock.h>
 #include <rdma/rdma_netlink.h>
 #include <linux/module.h>
 #include "core_priv.h"
 
 static DEFINE_MUTEX(rdma_nl_mutex);
-static struct sock *nls;
 static struct {
 	const struct rdma_nl_cbs   *cb_table;
 } rdma_nl_types[RDMA_NL_NUM_CLIENTS];
 
 bool rdma_nl_chk_listeners(unsigned int group)
 {
-	return netlink_has_listeners(nls, group);
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(&init_net);
+
+	return netlink_has_listeners(rnet->nl_sock, group);
 }
 EXPORT_SYMBOL(rdma_nl_chk_listeners);
 
@@ -73,13 +75,21 @@ static bool is_nl_msg_valid(unsigned int type, unsigned int op)
 	return (op < max_num_ops[type]) ? true : false;
 }
 
-static bool is_nl_valid(unsigned int type, unsigned int op)
+static bool
+is_nl_valid(const struct sk_buff *skb, unsigned int type, unsigned int op)
 {
 	const struct rdma_nl_cbs *cb_table;
 
 	if (!is_nl_msg_valid(type, op))
 		return false;
 
+	/*
+	 * Currently only NLDEV client is supporting netlink commands in
+	 * non init_net net namespace.
+	 */
+	if (sock_net(skb->sk) != &init_net && type != RDMA_NL_NLDEV)
+		return false;
+
 	if (!rdma_nl_types[type].cb_table) {
 		mutex_unlock(&rdma_nl_mutex);
 		request_module("rdma-netlink-subsys-%d", type);
@@ -161,7 +171,7 @@ static int rdma_nl_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh,
 	unsigned int op = RDMA_NL_GET_OP(type);
 	const struct rdma_nl_cbs *cb_table;
 
-	if (!is_nl_valid(index, op))
+	if (!is_nl_valid(skb, index, op))
 		return -EINVAL;
 
 	cb_table = rdma_nl_types[index].cb_table;
@@ -185,7 +195,7 @@ static int rdma_nl_rcv_msg(struct sk_buff *skb, struct nlmsghdr *nlh,
 			.dump = cb_table[op].dump,
 		};
 		if (c.dump)
-			return netlink_dump_start(nls, skb, nlh, &c);
+			return netlink_dump_start(skb->sk, skb, nlh, &c);
 		return -EINVAL;
 	}
 
@@ -258,52 +268,65 @@ static void rdma_nl_rcv(struct sk_buff *skb)
 	mutex_unlock(&rdma_nl_mutex);
 }
 
-int rdma_nl_unicast(struct sk_buff *skb, u32 pid)
+int rdma_nl_unicast(struct net *net, struct sk_buff *skb, u32 pid)
 {
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
 	int err;
 
-	err = netlink_unicast(nls, skb, pid, MSG_DONTWAIT);
+	err = netlink_unicast(rnet->nl_sock, skb, pid, MSG_DONTWAIT);
 	return (err < 0) ? err : 0;
 }
 EXPORT_SYMBOL(rdma_nl_unicast);
 
-int rdma_nl_unicast_wait(struct sk_buff *skb, __u32 pid)
+int rdma_nl_unicast_wait(struct net *net, struct sk_buff *skb, __u32 pid)
 {
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
 	int err;
 
-	err = netlink_unicast(nls, skb, pid, 0);
+	err = netlink_unicast(rnet->nl_sock, skb, pid, 0);
 	return (err < 0) ? err : 0;
 }
 EXPORT_SYMBOL(rdma_nl_unicast_wait);
 
-int rdma_nl_multicast(struct sk_buff *skb, unsigned int group, gfp_t flags)
+int rdma_nl_multicast(struct net *net, struct sk_buff *skb,
+		      unsigned int group, gfp_t flags)
 {
-	return nlmsg_multicast(nls, skb, 0, group, flags);
+	struct rdma_dev_net *rnet = rdma_net_to_dev_net(net);
+
+	return nlmsg_multicast(rnet->nl_sock, skb, 0, group, flags);
 }
 EXPORT_SYMBOL(rdma_nl_multicast);
 
-int __init rdma_nl_init(void)
-{
-	struct netlink_kernel_cfg cfg = {
-		.input	= rdma_nl_rcv,
-	};
-
-	nls = netlink_kernel_create(&init_net, NETLINK_RDMA, &cfg);
-	if (!nls)
-		return -ENOMEM;
-
-	nls->sk_sndtimeo = 10 * HZ;
-	return 0;
-}
-
 void rdma_nl_exit(void)
 {
 	int idx;
 
 	for (idx = 0; idx < RDMA_NL_NUM_CLIENTS; idx++)
-		rdma_nl_unregister(idx);
+		WARN(rdma_nl_types[idx].cb_table,
+		     "Netlink client %d wasn't released prior to unloading %s\n",
+		     idx, KBUILD_MODNAME);
+}
 
-	netlink_kernel_release(nls);
+int rdma_nl_net_init(struct rdma_dev_net *rnet)
+{
+	struct net *net = read_pnet(&rnet->net);
+	struct netlink_kernel_cfg cfg = {
+		.input	= rdma_nl_rcv,
+	};
+	struct sock *nls;
+
+	nls = netlink_kernel_create(net, NETLINK_RDMA, &cfg);
+	if (!nls)
+		return -ENOMEM;
+
+	nls->sk_sndtimeo = 10 * HZ;
+	rnet->nl_sock = nls;
+	return 0;
+}
+
+void rdma_nl_net_exit(struct rdma_dev_net *rnet)
+{
+	netlink_kernel_release(rnet->nl_sock);
 }
 
 MODULE_ALIAS_NET_PF_PROTO(PF_NETLINK, NETLINK_RDMA);

drivers/infiniband/core/nldev.c

@@ -831,7 +831,7 @@ static int nldev_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 	nlmsg_end(msg, nlh);
 
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_free:
 	nlmsg_free(msg);
@@ -971,7 +971,7 @@ static int nldev_port_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
 
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_free:
 	nlmsg_free(msg);
@@ -1073,7 +1073,7 @@ static int nldev_res_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_free:
 	nlmsg_free(msg);
@@ -1250,7 +1250,7 @@ static int res_get_common_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_free:
 	nlmsg_free(msg);
@@ -1595,7 +1595,7 @@ static int nldev_get_chardev(struct sk_buff *skb, struct nlmsghdr *nlh,
 	put_device(data.cdev);
 	if (ibdev)
 		ib_device_put(ibdev);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 out_data:
 	put_device(data.cdev);
@@ -1635,7 +1635,7 @@ static int nldev_sys_get_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 		return err;
 	}
 	nlmsg_end(msg, nlh);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 }
 
 static int nldev_set_sys_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
@@ -1733,7 +1733,7 @@ static int nldev_stat_set_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_fill:
 	rdma_counter_unbind_qpn(device, port, qpn, cntn);
@@ -1801,7 +1801,7 @@ static int nldev_stat_del_doit(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_fill:
 	rdma_counter_bind_qpn(device, port, qpn, cntn);
@@ -1892,7 +1892,7 @@ static int stat_get_doit_default_counter(struct sk_buff *skb,
 	mutex_unlock(&stats->lock);
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_table:
 	nla_nest_cancel(msg, table_attr);
@@ -1964,7 +1964,7 @@ static int stat_get_doit_qp(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 	nlmsg_end(msg, nlh);
 	ib_device_put(device);
-	return rdma_nl_unicast(msg, NETLINK_CB(skb).portid);
+	return rdma_nl_unicast(sock_net(skb->sk), msg, NETLINK_CB(skb).portid);
 
 err_msg:
 	nlmsg_free(msg);

drivers/infiniband/core/sa_query.c

@@ -860,7 +860,7 @@ static int ib_nl_send_msg(struct ib_sa_query *query, gfp_t gfp_mask)
 	/* Repair the nlmsg header length */
 	nlmsg_end(skb, nlh);
 
-	return rdma_nl_multicast(skb, RDMA_NL_GROUP_LS, gfp_mask);
+	return rdma_nl_multicast(&init_net, skb, RDMA_NL_GROUP_LS, gfp_mask);
 }
 
 static int ib_nl_make_request(struct ib_sa_query *query, gfp_t gfp_mask)

drivers/infiniband/core/sysfs.c

@@ -289,6 +289,24 @@ static ssize_t rate_show(struct ib_port *p, struct port_attribute *unused,
 		       ib_width_enum_to_int(attr.active_width), speed);
 }
 
+static const char *phys_state_to_str(enum ib_port_phys_state phys_state)
+{
+	static const char * phys_state_str[] = {
+		"<unknown>",
+		"Sleep",
+		"Polling",
+		"Disabled",
+		"PortConfigurationTraining",
+		"LinkUp",
+		"LinkErrorRecovery",
+		"Phy Test",
+	};
+
+	if (phys_state < ARRAY_SIZE(phys_state_str))
+		return phys_state_str[phys_state];
+	return "<unknown>";
+}
+
 static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused,
 			       char *buf)
 {
@@ -300,16 +318,8 @@ static ssize_t phys_state_show(struct ib_port *p, struct port_attribute *unused,
 	if (ret)
 		return ret;
 
-	switch (attr.phys_state) {
+	return sprintf(buf, "%d: %s\n", attr.phys_state,
-	case 1:  return sprintf(buf, "1: Sleep\n");
+		       phys_state_to_str(attr.phys_state));
-	case 2:  return sprintf(buf, "2: Polling\n");
-	case 3:  return sprintf(buf, "3: Disabled\n");
-	case 4:  return sprintf(buf, "4: PortConfigurationTraining\n");
-	case 5:  return sprintf(buf, "5: LinkUp\n");
-	case 6:  return sprintf(buf, "6: LinkErrorRecovery\n");
-	case 7:  return sprintf(buf, "7: Phy Test\n");
-	default: return sprintf(buf, "%d: <unknown>\n", attr.phys_state);
-	}
 }
 
 static ssize_t link_layer_show(struct ib_port *p, struct port_attribute *unused,

drivers/infiniband/core/umem_odp.c

@@ -218,7 +218,7 @@ static inline int ib_init_umem_odp(struct ib_umem_odp *umem_odp)
 		umem_odp->interval_tree.start =
 			ALIGN_DOWN(umem_odp->umem.address, page_size);
 		if (check_add_overflow(umem_odp->umem.address,
-				       umem_odp->umem.length,
+				       (unsigned long)umem_odp->umem.length,
 				       &umem_odp->interval_tree.last))
 			return -EOVERFLOW;
 		umem_odp->interval_tree.last =

drivers/infiniband/core/user_mad.c

@@ -1042,7 +1042,7 @@ static int ib_umad_close(struct inode *inode, struct file *filp)
 				ib_unregister_mad_agent(file->agent[i]);
 
 	mutex_unlock(&file->port->file_mutex);
-
+	mutex_destroy(&file->mutex);
 	kfree(file);
 	return 0;
 }

drivers/infiniband/core/uverbs_cmd.c

@@ -3479,7 +3479,8 @@ static int __uverbs_create_xsrq(struct uverbs_attr_bundle *attrs,
 
 err_copy:
 	ib_destroy_srq_user(srq, uverbs_get_cleared_udata(attrs));
-
+	/* It was released in ib_destroy_srq_user */
+	srq = NULL;
 err_free:
 	kfree(srq);
 err_put:

drivers/infiniband/core/uverbs_main.c

@@ -120,6 +120,8 @@ static void ib_uverbs_release_dev(struct device *device)
 
 	uverbs_destroy_api(dev->uapi);
 	cleanup_srcu_struct(&dev->disassociate_srcu);
+	mutex_destroy(&dev->lists_mutex);
+	mutex_destroy(&dev->xrcd_tree_mutex);
 	kfree(dev);
 }
 
@@ -212,6 +214,8 @@ void ib_uverbs_release_file(struct kref *ref)
 
 	if (file->disassociate_page)
 		__free_pages(file->disassociate_page, 0);
+	mutex_destroy(&file->umap_lock);
+	mutex_destroy(&file->ucontext_lock);
 	kfree(file);
 }
 

drivers/infiniband/core/verbs.c

@@ -2259,6 +2259,7 @@ int ib_dealloc_xrcd(struct ib_xrcd *xrcd, struct ib_udata *udata)
 		if (ret)
 			return ret;
 	}
+	mutex_destroy(&xrcd->tgt_qp_mutex);
 
 	return xrcd->device->ops.dealloc_xrcd(xrcd, udata);
 }

drivers/infiniband/hw/bnxt_re/hw_counters.c

@@ -74,7 +74,7 @@ static const char * const bnxt_re_stat_name[] = {
 	[BNXT_RE_SEQ_ERR_NAKS_RCVD]     = "seq_err_naks_rcvd",
 	[BNXT_RE_MAX_RETRY_EXCEEDED]    = "max_retry_exceeded",
 	[BNXT_RE_RNR_NAKS_RCVD]         = "rnr_naks_rcvd",
-	[BNXT_RE_MISSING_RESP]          = "missin_resp",
+	[BNXT_RE_MISSING_RESP]          = "missing_resp",
 	[BNXT_RE_UNRECOVERABLE_ERR]     = "unrecoverable_err",
 	[BNXT_RE_BAD_RESP_ERR]          = "bad_resp_err",
 	[BNXT_RE_LOCAL_QP_OP_ERR]       = "local_qp_op_err",

drivers/infiniband/hw/bnxt_re/ib_verbs.c

@@ -220,10 +220,10 @@ int bnxt_re_query_port(struct ib_device *ibdev, u8 port_num,
 
 	if (netif_running(rdev->netdev) && netif_carrier_ok(rdev->netdev)) {
 		port_attr->state = IB_PORT_ACTIVE;
-		port_attr->phys_state = 5;
+		port_attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	} else {
 		port_attr->state = IB_PORT_DOWN;
-		port_attr->phys_state = 3;
+		port_attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 	}
 	port_attr->max_mtu = IB_MTU_4096;
 	port_attr->active_mtu = iboe_get_mtu(rdev->netdev->mtu);
@@ -1398,7 +1398,7 @@ int bnxt_re_create_srq(struct ib_srq *ib_srq,
 			dev_err(rdev_to_dev(rdev), "SRQ copy to udata failed!");
 			bnxt_qplib_destroy_srq(&rdev->qplib_res,
 					       &srq->qplib_srq);
-			goto exit;
+			goto fail;
 		}
 	}
 	if (nq)

drivers/infiniband/hw/bnxt_re/main.c

@@ -1473,7 +1473,6 @@ static int bnxt_re_ib_reg(struct bnxt_re_dev *rdev)
 			 &rdev->active_width);
 	set_bit(BNXT_RE_FLAG_ISSUE_ROCE_STATS, &rdev->flags);
 	bnxt_re_dispatch_event(&rdev->ibdev, NULL, 1, IB_EVENT_PORT_ACTIVE);
-	bnxt_re_dispatch_event(&rdev->ibdev, NULL, 1, IB_EVENT_GID_CHANGE);
 
 	return 0;
 free_sctx:

drivers/infiniband/hw/cxgb3/iwch_provider.c

@@ -991,33 +991,8 @@ static int iwch_query_device(struct ib_device *ibdev, struct ib_device_attr *pro
 static int iwch_query_port(struct ib_device *ibdev,
 			   u8 port, struct ib_port_attr *props)
 {
-	struct iwch_dev *dev;
-	struct net_device *netdev;
-	struct in_device *inetdev;
-
 	pr_debug("%s ibdev %p\n", __func__, ibdev);
 
-	dev = to_iwch_dev(ibdev);
-	netdev = dev->rdev.port_info.lldevs[port-1];
-
-	/* props being zeroed by the caller, avoid zeroing it here */
-	props->max_mtu = IB_MTU_4096;
-	props->active_mtu = ib_mtu_int_to_enum(netdev->mtu);
-
-	if (!netif_carrier_ok(netdev))
-		props->state = IB_PORT_DOWN;
-	else {
-		inetdev = in_dev_get(netdev);
-		if (inetdev) {
-			if (inetdev->ifa_list)
-				props->state = IB_PORT_ACTIVE;
-			else
-				props->state = IB_PORT_INIT;
-			in_dev_put(inetdev);
-		} else
-			props->state = IB_PORT_INIT;
-	}
-
 	props->port_cap_flags =
 	    IB_PORT_CM_SUP |
 	    IB_PORT_SNMP_TUNNEL_SUP |
@@ -1273,8 +1248,24 @@ static const struct ib_device_ops iwch_dev_ops = {
 	INIT_RDMA_OBJ_SIZE(ib_ucontext, iwch_ucontext, ibucontext),
 };
 
+static int set_netdevs(struct ib_device *ib_dev, struct cxio_rdev *rdev)
+{
+	int ret;
+	int i;
+
+	for (i = 0; i < rdev->port_info.nports; i++) {
+		ret = ib_device_set_netdev(ib_dev, rdev->port_info.lldevs[i],
+					   i + 1);
+		if (ret)
+			return ret;
+	}
+	return 0;
+}
+
 int iwch_register_device(struct iwch_dev *dev)
 {
+	int err;
+
 	pr_debug("%s iwch_dev %p\n", __func__, dev);
 	memset(&dev->ibdev.node_guid, 0, sizeof(dev->ibdev.node_guid));
 	memcpy(&dev->ibdev.node_guid, dev->rdev.t3cdev_p->lldev->dev_addr, 6);
@@ -1315,6 +1306,10 @@ int iwch_register_device(struct iwch_dev *dev)
 
 	rdma_set_device_sysfs_group(&dev->ibdev, &iwch_attr_group);
 	ib_set_device_ops(&dev->ibdev, &iwch_dev_ops);
+	err = set_netdevs(&dev->ibdev, &dev->rdev);
+	if (err)
+		return err;
+
 	return ib_register_device(&dev->ibdev, "cxgb3_%d");
 }
 

drivers/infiniband/hw/cxgb4/provider.c

@@ -305,32 +305,8 @@ static int c4iw_query_device(struct ib_device *ibdev, struct ib_device_attr *pro
 static int c4iw_query_port(struct ib_device *ibdev, u8 port,
 			   struct ib_port_attr *props)
 {
-	struct c4iw_dev *dev;
-	struct net_device *netdev;
-	struct in_device *inetdev;
-
 	pr_debug("ibdev %p\n", ibdev);
 
-	dev = to_c4iw_dev(ibdev);
-	netdev = dev->rdev.lldi.ports[port-1];
-	/* props being zeroed by the caller, avoid zeroing it here */
-	props->max_mtu = IB_MTU_4096;
-	props->active_mtu = ib_mtu_int_to_enum(netdev->mtu);
-
-	if (!netif_carrier_ok(netdev))
-		props->state = IB_PORT_DOWN;
-	else {
-		inetdev = in_dev_get(netdev);
-		if (inetdev) {
-			if (inetdev->ifa_list)
-				props->state = IB_PORT_ACTIVE;
-			else
-				props->state = IB_PORT_INIT;
-			in_dev_put(inetdev);
-		} else
-			props->state = IB_PORT_INIT;
-	}
-
 	props->port_cap_flags =
 	    IB_PORT_CM_SUP |
 	    IB_PORT_SNMP_TUNNEL_SUP |

drivers/infiniband/hw/efa/efa.h

@@ -156,5 +156,8 @@ int efa_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
 		  int qp_attr_mask, struct ib_udata *udata);
 enum rdma_link_layer efa_port_link_layer(struct ib_device *ibdev,
 					 u8 port_num);
+struct rdma_hw_stats *efa_alloc_hw_stats(struct ib_device *ibdev, u8 port_num);
+int efa_get_hw_stats(struct ib_device *ibdev, struct rdma_hw_stats *stats,
+		     u8 port_num, int index);
 
 #endif /* _EFA_H_ */

drivers/infiniband/hw/efa/efa_com.c

@@ -109,17 +109,19 @@ static u32 efa_com_reg_read32(struct efa_com_dev *edev, u16 offset)
 	} while (time_is_after_jiffies(exp_time));
 
 	if (read_resp->req_id != mmio_read->seq_num) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Reading register timed out. expected: req id[%u] offset[%#x] actual: req id[%u] offset[%#x]\n",
+			edev->efa_dev,
-			  mmio_read->seq_num, offset, read_resp->req_id,
+			"Reading register timed out. expected: req id[%u] offset[%#x] actual: req id[%u] offset[%#x]\n",
-			  read_resp->reg_off);
+			mmio_read->seq_num, offset, read_resp->req_id,
+			read_resp->reg_off);
 		err = EFA_MMIO_READ_INVALID;
 		goto out;
 	}
 
 	if (read_resp->reg_off != offset) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Reading register failed: wrong offset provided\n");
+			edev->efa_dev,
+			"Reading register failed: wrong offset provided\n");
 		err = EFA_MMIO_READ_INVALID;
 		goto out;
 	}
@@ -293,9 +295,10 @@ static struct efa_comp_ctx *efa_com_get_comp_ctx(struct efa_com_admin_queue *aq,
 	u16 ctx_id = cmd_id & (aq->depth - 1);
 
 	if (aq->comp_ctx[ctx_id].occupied && capture) {
-		ibdev_err(aq->efa_dev,
+		ibdev_err_ratelimited(
-			  "Completion context for command_id %#x is occupied\n",
+			aq->efa_dev,
-			  cmd_id);
+			"Completion context for command_id %#x is occupied\n",
+			cmd_id);
 		return NULL;
 	}
 
@@ -401,7 +404,7 @@ static struct efa_comp_ctx *efa_com_submit_admin_cmd(struct efa_com_admin_queue
 
 	spin_lock(&aq->sq.lock);
 	if (!test_bit(EFA_AQ_STATE_RUNNING_BIT, &aq->state)) {
-		ibdev_err(aq->efa_dev, "Admin queue is closed\n");
+		ibdev_err_ratelimited(aq->efa_dev, "Admin queue is closed\n");
 		spin_unlock(&aq->sq.lock);
 		return ERR_PTR(-ENODEV);
 	}
@@ -519,8 +522,9 @@ static int efa_com_wait_and_process_admin_cq_polling(struct efa_comp_ctx *comp_c
 			break;
 
 		if (time_is_before_jiffies(timeout)) {
-			ibdev_err(aq->efa_dev,
+			ibdev_err_ratelimited(
-				  "Wait for completion (polling) timeout\n");
+				aq->efa_dev,
+				"Wait for completion (polling) timeout\n");
 			/* EFA didn't have any completion */
 			atomic64_inc(&aq->stats.no_completion);
 
@@ -561,17 +565,19 @@ static int efa_com_wait_and_process_admin_cq_interrupts(struct efa_comp_ctx *com
 		atomic64_inc(&aq->stats.no_completion);
 
 		if (comp_ctx->status == EFA_CMD_COMPLETED)
-			ibdev_err(aq->efa_dev,
+			ibdev_err_ratelimited(
-				  "The device sent a completion but the driver didn't receive any MSI-X interrupt for admin cmd %s(%d) status %d (ctx: 0x%p, sq producer: %d, sq consumer: %d, cq consumer: %d)\n",
+				aq->efa_dev,
-				  efa_com_cmd_str(comp_ctx->cmd_opcode),
+				"The device sent a completion but the driver didn't receive any MSI-X interrupt for admin cmd %s(%d) status %d (ctx: 0x%p, sq producer: %d, sq consumer: %d, cq consumer: %d)\n",
-				  comp_ctx->cmd_opcode, comp_ctx->status,
+				efa_com_cmd_str(comp_ctx->cmd_opcode),
-				  comp_ctx, aq->sq.pc, aq->sq.cc, aq->cq.cc);
+				comp_ctx->cmd_opcode, comp_ctx->status,
+				comp_ctx, aq->sq.pc, aq->sq.cc, aq->cq.cc);
 		else
-			ibdev_err(aq->efa_dev,
+			ibdev_err_ratelimited(
-				  "The device didn't send any completion for admin cmd %s(%d) status %d (ctx 0x%p, sq producer: %d, sq consumer: %d, cq consumer: %d)\n",
+				aq->efa_dev,
-				  efa_com_cmd_str(comp_ctx->cmd_opcode),
+				"The device didn't send any completion for admin cmd %s(%d) status %d (ctx 0x%p, sq producer: %d, sq consumer: %d, cq consumer: %d)\n",
-				  comp_ctx->cmd_opcode, comp_ctx->status,
+				efa_com_cmd_str(comp_ctx->cmd_opcode),
-				  comp_ctx, aq->sq.pc, aq->sq.cc, aq->cq.cc);
+				comp_ctx->cmd_opcode, comp_ctx->status,
+				comp_ctx, aq->sq.pc, aq->sq.cc, aq->cq.cc);
 
 		clear_bit(EFA_AQ_STATE_RUNNING_BIT, &aq->state);
 		err = -ETIME;
@@ -633,10 +639,11 @@ int efa_com_cmd_exec(struct efa_com_admin_queue *aq,
 		  cmd->aq_common_descriptor.opcode);
 	comp_ctx = efa_com_submit_admin_cmd(aq, cmd, cmd_size, comp, comp_size);
 	if (IS_ERR(comp_ctx)) {
-		ibdev_err(aq->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to submit command %s (opcode %u) err %ld\n",
+			aq->efa_dev,
-			  efa_com_cmd_str(cmd->aq_common_descriptor.opcode),
+			"Failed to submit command %s (opcode %u) err %ld\n",
-			  cmd->aq_common_descriptor.opcode, PTR_ERR(comp_ctx));
+			efa_com_cmd_str(cmd->aq_common_descriptor.opcode),
+			cmd->aq_common_descriptor.opcode, PTR_ERR(comp_ctx));
 
 		up(&aq->avail_cmds);
 		return PTR_ERR(comp_ctx);
@@ -644,11 +651,12 @@ int efa_com_cmd_exec(struct efa_com_admin_queue *aq,
 
 	err = efa_com_wait_and_process_admin_cq(comp_ctx, aq);
 	if (err)
-		ibdev_err(aq->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to process command %s (opcode %u) comp_status %d err %d\n",
+			aq->efa_dev,
-			  efa_com_cmd_str(cmd->aq_common_descriptor.opcode),
+			"Failed to process command %s (opcode %u) comp_status %d err %d\n",
-			  cmd->aq_common_descriptor.opcode,
+			efa_com_cmd_str(cmd->aq_common_descriptor.opcode),
-			  comp_ctx->comp_status, err);
+			cmd->aq_common_descriptor.opcode, comp_ctx->comp_status,
+			err);
 
 	up(&aq->avail_cmds);
 

drivers/infiniband/hw/efa/efa_com_cmd.c

@@ -44,7 +44,8 @@ int efa_com_create_qp(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to create qp [%d]\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to create qp [%d]\n", err);
 		return err;
 	}
 
@@ -82,9 +83,10 @@ int efa_com_modify_qp(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to modify qp-%u modify_mask[%#x] [%d]\n",
+			edev->efa_dev,
-			  cmd.qp_handle, cmd.modify_mask, err);
+			"Failed to modify qp-%u modify_mask[%#x] [%d]\n",
+			cmd.qp_handle, cmd.modify_mask, err);
 		return err;
 	}
 
@@ -109,8 +111,9 @@ int efa_com_query_qp(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to query qp-%u [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  cmd.qp_handle, err);
+				      "Failed to query qp-%u [%d]\n",
+				      cmd.qp_handle, err);
 		return err;
 	}
 
@@ -139,8 +142,9 @@ int efa_com_destroy_qp(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to destroy qp-%u [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  qp_cmd.qp_handle, err);
+				      "Failed to destroy qp-%u [%d]\n",
+				      qp_cmd.qp_handle, err);
 		return err;
 	}
 
@@ -173,7 +177,8 @@ int efa_com_create_cq(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to create cq[%d]\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to create cq[%d]\n", err);
 		return err;
 	}
 
@@ -201,8 +206,9 @@ int efa_com_destroy_cq(struct efa_com_dev *edev,
 			       sizeof(destroy_resp));
 
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to destroy CQ-%u [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  params->cq_idx, err);
+				      "Failed to destroy CQ-%u [%d]\n",
+				      params->cq_idx, err);
 		return err;
 	}
 
@@ -250,7 +256,8 @@ int efa_com_register_mr(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to register mr [%d]\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to register mr [%d]\n", err);
 		return err;
 	}
 
@@ -277,9 +284,9 @@ int efa_com_dereg_mr(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(edev->efa_dev,
-			  "Failed to de-register mr(lkey-%u) [%d]\n",
+				      "Failed to de-register mr(lkey-%u) [%d]\n",
-			  mr_cmd.l_key, err);
+				      mr_cmd.l_key, err);
 		return err;
 	}
 
@@ -306,8 +313,9 @@ int efa_com_create_ah(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to create ah for %pI6 [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  ah_cmd.dest_addr, err);
+				      "Failed to create ah for %pI6 [%d]\n",
+				      ah_cmd.dest_addr, err);
 		return err;
 	}
 
@@ -334,8 +342,9 @@ int efa_com_destroy_ah(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&cmd_completion,
 			       sizeof(cmd_completion));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to destroy ah-%d pd-%d [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  ah_cmd.ah, ah_cmd.pd, err);
+				      "Failed to destroy ah-%d pd-%d [%d]\n",
+				      ah_cmd.ah, ah_cmd.pd, err);
 		return err;
 	}
 
@@ -367,8 +376,9 @@ static int efa_com_get_feature_ex(struct efa_com_dev *edev,
 	int err;
 
 	if (!efa_com_check_supported_feature_id(edev, feature_id)) {
-		ibdev_err(edev->efa_dev, "Feature %d isn't supported\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  feature_id);
+				      "Feature %d isn't supported\n",
+				      feature_id);
 		return -EOPNOTSUPP;
 	}
 
@@ -396,9 +406,10 @@ static int efa_com_get_feature_ex(struct efa_com_dev *edev,
 			       sizeof(*get_resp));
 
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to submit get_feature command %d [%d]\n",
+			edev->efa_dev,
-			  feature_id, err);
+			"Failed to submit get_feature command %d [%d]\n",
+			feature_id, err);
 		return err;
 	}
 
@@ -421,8 +432,9 @@ int efa_com_get_network_attr(struct efa_com_dev *edev,
 	err = efa_com_get_feature(edev, &resp,
 				  EFA_ADMIN_NETWORK_ATTR);
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(edev->efa_dev,
-			  "Failed to get network attributes %d\n", err);
+				      "Failed to get network attributes %d\n",
+				      err);
 		return err;
 	}
 
@@ -441,8 +453,9 @@ int efa_com_get_device_attr(struct efa_com_dev *edev,
 
 	err = efa_com_get_feature(edev, &resp, EFA_ADMIN_DEVICE_ATTR);
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to get device attributes %d\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  err);
+				      "Failed to get device attributes %d\n",
+				      err);
 		return err;
 	}
 
@@ -456,9 +469,10 @@ int efa_com_get_device_attr(struct efa_com_dev *edev,
 	result->db_bar = resp.u.device_attr.db_bar;
 
 	if (result->admin_api_version < 1) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to get device attr api version [%u < 1]\n",
+			edev->efa_dev,
-			  result->admin_api_version);
+			"Failed to get device attr api version [%u < 1]\n",
+			result->admin_api_version);
 		return -EINVAL;
 	}
 
@@ -466,8 +480,9 @@ int efa_com_get_device_attr(struct efa_com_dev *edev,
 	err = efa_com_get_feature(edev, &resp,
 				  EFA_ADMIN_QUEUE_ATTR);
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(edev->efa_dev,
-			  "Failed to get network attributes %d\n", err);
+				      "Failed to get queue attributes %d\n",
+				      err);
 		return err;
 	}
 
@@ -497,7 +512,8 @@ int efa_com_get_hw_hints(struct efa_com_dev *edev,
 
 	err = efa_com_get_feature(edev, &resp, EFA_ADMIN_HW_HINTS);
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to get hw hints %d\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to get hw hints %d\n", err);
 		return err;
 	}
 
@@ -520,8 +536,9 @@ static int efa_com_set_feature_ex(struct efa_com_dev *edev,
 	int err;
 
 	if (!efa_com_check_supported_feature_id(edev, feature_id)) {
-		ibdev_err(edev->efa_dev, "Feature %d isn't supported\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  feature_id);
+				      "Feature %d isn't supported\n",
+				      feature_id);
 		return -EOPNOTSUPP;
 	}
 
@@ -545,9 +562,10 @@ static int efa_com_set_feature_ex(struct efa_com_dev *edev,
 			       sizeof(*set_resp));
 
 	if (err) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Failed to submit set_feature command %d error: %d\n",
+			edev->efa_dev,
-			  feature_id, err);
+			"Failed to submit set_feature command %d error: %d\n",
+			feature_id, err);
 		return err;
 	}
 
@@ -574,8 +592,9 @@ int efa_com_set_aenq_config(struct efa_com_dev *edev, u32 groups)
 
 	err = efa_com_get_feature(edev, &get_resp, EFA_ADMIN_AENQ_CONFIG);
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to get aenq attributes: %d\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  err);
+				      "Failed to get aenq attributes: %d\n",
+				      err);
 		return err;
 	}
 
@@ -585,9 +604,10 @@ int efa_com_set_aenq_config(struct efa_com_dev *edev, u32 groups)
 		  get_resp.u.aenq.enabled_groups);
 
 	if ((get_resp.u.aenq.supported_groups & groups) != groups) {
-		ibdev_err(edev->efa_dev,
+		ibdev_err_ratelimited(
-			  "Trying to set unsupported aenq groups[%#x] supported[%#x]\n",
+			edev->efa_dev,
-			  groups, get_resp.u.aenq.supported_groups);
+			"Trying to set unsupported aenq groups[%#x] supported[%#x]\n",
+			groups, get_resp.u.aenq.supported_groups);
 		return -EOPNOTSUPP;
 	}
 
@@ -595,8 +615,9 @@ int efa_com_set_aenq_config(struct efa_com_dev *edev, u32 groups)
 	err = efa_com_set_feature(edev, &set_resp, &cmd,
 				  EFA_ADMIN_AENQ_CONFIG);
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to set aenq attributes: %d\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  err);
+				      "Failed to set aenq attributes: %d\n",
+				      err);
 		return err;
 	}
 
@@ -619,7 +640,8 @@ int efa_com_alloc_pd(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to allocate pd[%d]\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to allocate pd[%d]\n", err);
 		return err;
 	}
 
@@ -645,8 +667,9 @@ int efa_com_dealloc_pd(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to deallocate pd-%u [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  cmd.pd, err);
+				      "Failed to deallocate pd-%u [%d]\n",
+				      cmd.pd, err);
 		return err;
 	}
 
@@ -669,7 +692,8 @@ int efa_com_alloc_uar(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to allocate uar[%d]\n", err);
+		ibdev_err_ratelimited(edev->efa_dev,
+				      "Failed to allocate uar[%d]\n", err);
 		return err;
 	}
 
@@ -695,10 +719,47 @@ int efa_com_dealloc_uar(struct efa_com_dev *edev,
 			       (struct efa_admin_acq_entry *)&resp,
 			       sizeof(resp));
 	if (err) {
-		ibdev_err(edev->efa_dev, "Failed to deallocate uar-%u [%d]\n",
+		ibdev_err_ratelimited(edev->efa_dev,
-			  cmd.uar, err);
+				      "Failed to deallocate uar-%u [%d]\n",
+				      cmd.uar, err);
 		return err;
 	}
 
 	return 0;
 }
+
+int efa_com_get_stats(struct efa_com_dev *edev,
+		      struct efa_com_get_stats_params *params,
+		      union efa_com_get_stats_result *result)
+{
+	struct efa_com_admin_queue *aq = &edev->aq;
+	struct efa_admin_aq_get_stats_cmd cmd = {};
+	struct efa_admin_acq_get_stats_resp resp;
+	int err;
+
+	cmd.aq_common_descriptor.opcode = EFA_ADMIN_GET_STATS;
+	cmd.type = params->type;
+	cmd.scope = params->scope;
+	cmd.scope_modifier = params->scope_modifier;
+
+	err = efa_com_cmd_exec(aq,
+			       (struct efa_admin_aq_entry *)&cmd,
+			       sizeof(cmd),
+			       (struct efa_admin_acq_entry *)&resp,
+			       sizeof(resp));
+	if (err) {
+		ibdev_err_ratelimited(
+			edev->efa_dev,
+			"Failed to get stats type-%u scope-%u.%u [%d]\n",
+			cmd.type, cmd.scope, cmd.scope_modifier, err);
+		return err;
+	}
+
+	result->basic_stats.tx_bytes = resp.basic_stats.tx_bytes;
+	result->basic_stats.tx_pkts = resp.basic_stats.tx_pkts;
+	result->basic_stats.rx_bytes = resp.basic_stats.rx_bytes;
+	result->basic_stats.rx_pkts = resp.basic_stats.rx_pkts;
+	result->basic_stats.rx_drops = resp.basic_stats.rx_drops;
+
+	return 0;
+}

drivers/infiniband/hw/efa/efa_com_cmd.h

@@ -225,6 +225,26 @@ struct efa_com_dealloc_uar_params {
 	u16 uarn;
 };
 
+struct efa_com_get_stats_params {
+	/* see enum efa_admin_get_stats_type */
+	u8 type;
+	/* see enum efa_admin_get_stats_scope */
+	u8 scope;
+	u16 scope_modifier;
+};
+
+struct efa_com_basic_stats {
+	u64 tx_bytes;
+	u64 tx_pkts;
+	u64 rx_bytes;
+	u64 rx_pkts;
+	u64 rx_drops;
+};
+
+union efa_com_get_stats_result {
+	struct efa_com_basic_stats basic_stats;
+};
+
 void efa_com_set_dma_addr(dma_addr_t addr, u32 *addr_high, u32 *addr_low);
 int efa_com_create_qp(struct efa_com_dev *edev,
 		      struct efa_com_create_qp_params *params,
@@ -266,5 +286,8 @@ int efa_com_alloc_uar(struct efa_com_dev *edev,
 		      struct efa_com_alloc_uar_result *result);
 int efa_com_dealloc_uar(struct efa_com_dev *edev,
 			struct efa_com_dealloc_uar_params *params);
+int efa_com_get_stats(struct efa_com_dev *edev,
+		      struct efa_com_get_stats_params *params,
+		      union efa_com_get_stats_result *result);
 
 #endif /* _EFA_COM_CMD_H_ */

drivers/infiniband/hw/efa/efa_main.c

@@ -201,6 +201,7 @@ static const struct ib_device_ops efa_dev_ops = {
 	.driver_id = RDMA_DRIVER_EFA,
 	.uverbs_abi_ver = EFA_UVERBS_ABI_VERSION,
 
+	.alloc_hw_stats = efa_alloc_hw_stats,
 	.alloc_pd = efa_alloc_pd,
 	.alloc_ucontext = efa_alloc_ucontext,
 	.create_ah = efa_create_ah,
@@ -212,6 +213,7 @@ static const struct ib_device_ops efa_dev_ops = {
 	.destroy_ah = efa_destroy_ah,
 	.destroy_cq = efa_destroy_cq,
 	.destroy_qp = efa_destroy_qp,
+	.get_hw_stats = efa_get_hw_stats,
 	.get_link_layer = efa_port_link_layer,
 	.get_port_immutable = efa_get_port_immutable,
 	.mmap = efa_mmap,

drivers/infiniband/hw/efa/efa_verbs.c

@@ -41,6 +41,33 @@ static inline u64 get_mmap_key(const struct efa_mmap_entry *efa)
 	       ((u64)efa->mmap_page << PAGE_SHIFT);
 }
 
+#define EFA_DEFINE_STATS(op) \
+	op(EFA_TX_BYTES, "tx_bytes") \
+	op(EFA_TX_PKTS, "tx_pkts") \
+	op(EFA_RX_BYTES, "rx_bytes") \
+	op(EFA_RX_PKTS, "rx_pkts") \
+	op(EFA_RX_DROPS, "rx_drops") \
+	op(EFA_SUBMITTED_CMDS, "submitted_cmds") \
+	op(EFA_COMPLETED_CMDS, "completed_cmds") \
+	op(EFA_NO_COMPLETION_CMDS, "no_completion_cmds") \
+	op(EFA_KEEP_ALIVE_RCVD, "keep_alive_rcvd") \
+	op(EFA_ALLOC_PD_ERR, "alloc_pd_err") \
+	op(EFA_CREATE_QP_ERR, "create_qp_err") \
+	op(EFA_REG_MR_ERR, "reg_mr_err") \
+	op(EFA_ALLOC_UCONTEXT_ERR, "alloc_ucontext_err") \
+	op(EFA_CREATE_AH_ERR, "create_ah_err")
+
+#define EFA_STATS_ENUM(ename, name) ename,
+#define EFA_STATS_STR(ename, name) [ename] = name,
+
+enum efa_hw_stats {
+	EFA_DEFINE_STATS(EFA_STATS_ENUM)
+};
+
+static const char *const efa_stats_names[] = {
+	EFA_DEFINE_STATS(EFA_STATS_STR)
+};
+
 #define EFA_CHUNK_PAYLOAD_SHIFT       12
 #define EFA_CHUNK_PAYLOAD_SIZE        BIT(EFA_CHUNK_PAYLOAD_SHIFT)
 #define EFA_CHUNK_PAYLOAD_PTR_SIZE    8
@@ -121,7 +148,7 @@ static inline struct efa_ah *to_eah(struct ib_ah *ibah)
 }
 
 #define field_avail(x, fld, sz) (offsetof(typeof(x), fld) + \
-				 sizeof(((typeof(x) *)0)->fld) <= (sz))
+				 FIELD_SIZEOF(typeof(x), fld) <= (sz))
 
 #define is_reserved_cleared(reserved) \
 	!memchr_inv(reserved, 0, sizeof(reserved))
@@ -306,7 +333,7 @@ int efa_query_port(struct ib_device *ibdev, u8 port,
 	props->lmc = 1;
 
 	props->state = IB_PORT_ACTIVE;
-	props->phys_state = 5;
+	props->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	props->gid_tbl_len = 1;
 	props->pkey_tbl_len = 1;
 	props->active_speed = IB_SPEED_EDR;
@@ -1473,14 +1500,12 @@ int efa_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata)
 
 	ibdev_dbg(&dev->ibdev, "Deregister mr[%d]\n", ibmr->lkey);
 
-	if (mr->umem) {
+	params.l_key = mr->ibmr.lkey;
-		params.l_key = mr->ibmr.lkey;
+	err = efa_com_dereg_mr(&dev->edev, &params);
-		err = efa_com_dereg_mr(&dev->edev, &params);
+	if (err)
-		if (err)
+		return err;
-			return err;
-	}
-	ib_umem_release(mr->umem);
 
+	ib_umem_release(mr->umem);
 	kfree(mr);
 
 	return 0;
@@ -1727,6 +1752,54 @@ void efa_destroy_ah(struct ib_ah *ibah, u32 flags)
 	efa_ah_destroy(dev, ah);
 }
 
+struct rdma_hw_stats *efa_alloc_hw_stats(struct ib_device *ibdev, u8 port_num)
+{
+	return rdma_alloc_hw_stats_struct(efa_stats_names,
+					  ARRAY_SIZE(efa_stats_names),
+					  RDMA_HW_STATS_DEFAULT_LIFESPAN);
+}
+
+int efa_get_hw_stats(struct ib_device *ibdev, struct rdma_hw_stats *stats,
+		     u8 port_num, int index)
+{
+	struct efa_com_get_stats_params params = {};
+	union efa_com_get_stats_result result;
+	struct efa_dev *dev = to_edev(ibdev);
+	struct efa_com_basic_stats *bs;
+	struct efa_com_stats_admin *as;
+	struct efa_stats *s;
+	int err;
+
+	params.type = EFA_ADMIN_GET_STATS_TYPE_BASIC;
+	params.scope = EFA_ADMIN_GET_STATS_SCOPE_ALL;
+
+	err = efa_com_get_stats(&dev->edev, &params, &result);
+	if (err)
+		return err;
+
+	bs = &result.basic_stats;
+	stats->value[EFA_TX_BYTES] = bs->tx_bytes;
+	stats->value[EFA_TX_PKTS] = bs->tx_pkts;
+	stats->value[EFA_RX_BYTES] = bs->rx_bytes;
+	stats->value[EFA_RX_PKTS] = bs->rx_pkts;
+	stats->value[EFA_RX_DROPS] = bs->rx_drops;
+
+	as = &dev->edev.aq.stats;
+	stats->value[EFA_SUBMITTED_CMDS] = atomic64_read(&as->submitted_cmd);
+	stats->value[EFA_COMPLETED_CMDS] = atomic64_read(&as->completed_cmd);
+	stats->value[EFA_NO_COMPLETION_CMDS] = atomic64_read(&as->no_completion);
+
+	s = &dev->stats;
+	stats->value[EFA_KEEP_ALIVE_RCVD] = atomic64_read(&s->keep_alive_rcvd);
+	stats->value[EFA_ALLOC_PD_ERR] = atomic64_read(&s->sw_stats.alloc_pd_err);
+	stats->value[EFA_CREATE_QP_ERR] = atomic64_read(&s->sw_stats.create_qp_err);
+	stats->value[EFA_REG_MR_ERR] = atomic64_read(&s->sw_stats.reg_mr_err);
+	stats->value[EFA_ALLOC_UCONTEXT_ERR] = atomic64_read(&s->sw_stats.alloc_ucontext_err);
+	stats->value[EFA_CREATE_AH_ERR] = atomic64_read(&s->sw_stats.create_ah_err);
+
+	return ARRAY_SIZE(efa_stats_names);
+}
+
 enum rdma_link_layer efa_port_link_layer(struct ib_device *ibdev,
 					 u8 port_num)
 {

drivers/infiniband/hw/hfi1/chip.c

@@ -4101,6 +4101,7 @@ def_access_ibp_counter(rc_dupreq);
 def_access_ibp_counter(rdma_seq);
 def_access_ibp_counter(unaligned);
 def_access_ibp_counter(seq_naks);
+def_access_ibp_counter(rc_crwaits);
 
 static struct cntr_entry dev_cntrs[DEV_CNTR_LAST] = {
 [C_RCV_OVF] = RXE32_DEV_CNTR_ELEM(RcvOverflow, RCV_BUF_OVFL_CNT, CNTR_SYNTH),
@@ -5119,6 +5120,7 @@ static struct cntr_entry port_cntrs[PORT_CNTR_LAST] = {
 [C_SW_IBP_RDMA_SEQ] = SW_IBP_CNTR(RdmaSeq, rdma_seq),
 [C_SW_IBP_UNALIGNED] = SW_IBP_CNTR(Unaligned, unaligned),
 [C_SW_IBP_SEQ_NAK] = SW_IBP_CNTR(SeqNak, seq_naks),
+[C_SW_IBP_RC_CRWAITS] = SW_IBP_CNTR(RcCrWait, rc_crwaits),
 [C_SW_CPU_RC_ACKS] = CNTR_ELEM("RcAcks", 0, 0, CNTR_NORMAL,
 			       access_sw_cpu_rc_acks),
 [C_SW_CPU_RC_QACKS] = CNTR_ELEM("RcQacks", 0, 0, CNTR_NORMAL,

drivers/infiniband/hw/hfi1/chip.h

@@ -1245,6 +1245,7 @@ enum {
 	C_SW_IBP_RDMA_SEQ,
 	C_SW_IBP_UNALIGNED,
 	C_SW_IBP_SEQ_NAK,
+	C_SW_IBP_RC_CRWAITS,
 	C_SW_CPU_RC_ACKS,
 	C_SW_CPU_RC_QACKS,
 	C_SW_CPU_RC_DELAYED_COMP,

drivers/infiniband/hw/hfi1/mad.c

@@ -2326,7 +2326,7 @@ struct opa_port_status_req {
 	__be32 vl_select_mask;
 };
 
-#define VL_MASK_ALL		0x000080ff
+#define VL_MASK_ALL		0x00000000000080ffUL
 
 struct opa_port_status_rsp {
 	__u8 port_num;
@@ -2625,15 +2625,14 @@ static int pma_get_opa_classportinfo(struct opa_pma_mad *pmp,
 }
 
 static void a0_portstatus(struct hfi1_pportdata *ppd,
-			  struct opa_port_status_rsp *rsp, u32 vl_select_mask)
+			  struct opa_port_status_rsp *rsp)
 {
 	if (!is_bx(ppd->dd)) {
 		unsigned long vl;
 		u64 sum_vl_xmit_wait = 0;
-		u32 vl_all_mask = VL_MASK_ALL;
+		unsigned long vl_all_mask = VL_MASK_ALL;
 
-		for_each_set_bit(vl, (unsigned long *)&(vl_all_mask),
+		for_each_set_bit(vl, &vl_all_mask, BITS_PER_LONG) {
-				 8 * sizeof(vl_all_mask)) {
 			u64 tmp = sum_vl_xmit_wait +
 				  read_port_cntr(ppd, C_TX_WAIT_VL,
 						 idx_from_vl(vl));
@@ -2730,12 +2729,12 @@ static int pma_get_opa_portstatus(struct opa_pma_mad *pmp,
 		(struct opa_port_status_req *)pmp->data;
 	struct hfi1_devdata *dd = dd_from_ibdev(ibdev);
 	struct opa_port_status_rsp *rsp;
-	u32 vl_select_mask = be32_to_cpu(req->vl_select_mask);
+	unsigned long vl_select_mask = be32_to_cpu(req->vl_select_mask);
 	unsigned long vl;
 	size_t response_data_size;
 	u32 nports = be32_to_cpu(pmp->mad_hdr.attr_mod) >> 24;
 	u8 port_num = req->port_num;
-	u8 num_vls = hweight32(vl_select_mask);
+	u8 num_vls = hweight64(vl_select_mask);
 	struct _vls_pctrs *vlinfo;
 	struct hfi1_ibport *ibp = to_iport(ibdev, port);
 	struct hfi1_pportdata *ppd = ppd_from_ibp(ibp);
@@ -2770,7 +2769,7 @@ static int pma_get_opa_portstatus(struct opa_pma_mad *pmp,
 
 	hfi1_read_link_quality(dd, &rsp->link_quality_indicator);
 
-	rsp->vl_select_mask = cpu_to_be32(vl_select_mask);
+	rsp->vl_select_mask = cpu_to_be32((u32)vl_select_mask);
 	rsp->port_xmit_data = cpu_to_be64(read_dev_cntr(dd, C_DC_XMIT_FLITS,
 					  CNTR_INVALID_VL));
 	rsp->port_rcv_data = cpu_to_be64(read_dev_cntr(dd, C_DC_RCV_FLITS,
@@ -2841,8 +2840,7 @@ static int pma_get_opa_portstatus(struct opa_pma_mad *pmp,
 	 * So in the for_each_set_bit() loop below, we don't need
 	 * any additional checks for vl.
 	 */
-	for_each_set_bit(vl, (unsigned long *)&(vl_select_mask),
+	for_each_set_bit(vl, &vl_select_mask, BITS_PER_LONG) {
-			 8 * sizeof(vl_select_mask)) {
 		memset(vlinfo, 0, sizeof(*vlinfo));
 
 		tmp = read_dev_cntr(dd, C_DC_RX_FLIT_VL, idx_from_vl(vl));
@@ -2883,7 +2881,7 @@ static int pma_get_opa_portstatus(struct opa_pma_mad *pmp,
 		vfi++;
 	}
 
-	a0_portstatus(ppd, rsp, vl_select_mask);
+	a0_portstatus(ppd, rsp);
 
 	if (resp_len)
 		*resp_len += response_data_size;
@@ -2930,16 +2928,14 @@ static u64 get_error_counter_summary(struct ib_device *ibdev, u8 port,
 	return error_counter_summary;
 }
 
-static void a0_datacounters(struct hfi1_pportdata *ppd, struct _port_dctrs *rsp,
+static void a0_datacounters(struct hfi1_pportdata *ppd, struct _port_dctrs *rsp)
-			    u32 vl_select_mask)
 {
 	if (!is_bx(ppd->dd)) {
 		unsigned long vl;
 		u64 sum_vl_xmit_wait = 0;
-		u32 vl_all_mask = VL_MASK_ALL;
+		unsigned long vl_all_mask = VL_MASK_ALL;
 
-		for_each_set_bit(vl, (unsigned long *)&(vl_all_mask),
+		for_each_set_bit(vl, &vl_all_mask, BITS_PER_LONG) {
-				 8 * sizeof(vl_all_mask)) {
 			u64 tmp = sum_vl_xmit_wait +
 				  read_port_cntr(ppd, C_TX_WAIT_VL,
 						 idx_from_vl(vl));
@@ -2994,7 +2990,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp,
 	u64 port_mask;
 	u8 port_num;
 	unsigned long vl;
-	u32 vl_select_mask;
+	unsigned long vl_select_mask;
 	int vfi;
 	u16 link_width;
 	u16 link_speed;
@@ -3071,8 +3067,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp,
 	 * So in the for_each_set_bit() loop below, we don't need
 	 * any additional checks for vl.
 	 */
-	for_each_set_bit(vl, (unsigned long *)&(vl_select_mask),
+	for_each_set_bit(vl, &vl_select_mask, BITS_PER_LONG) {
-			 8 * sizeof(req->vl_select_mask)) {
 		memset(vlinfo, 0, sizeof(*vlinfo));
 
 		rsp->vls[vfi].port_vl_xmit_data =
@@ -3120,7 +3115,7 @@ static int pma_get_opa_datacounters(struct opa_pma_mad *pmp,
 		vfi++;
 	}
 
-	a0_datacounters(ppd, rsp, vl_select_mask);
+	a0_datacounters(ppd, rsp);
 
 	if (resp_len)
 		*resp_len += response_data_size;
@@ -3215,7 +3210,7 @@ static int pma_get_opa_porterrors(struct opa_pma_mad *pmp,
 	struct _vls_ectrs *vlinfo;
 	unsigned long vl;
 	u64 port_mask, tmp;
-	u32 vl_select_mask;
+	unsigned long vl_select_mask;
 	int vfi;
 
 	req = (struct opa_port_error_counters64_msg *)pmp->data;
@@ -3273,8 +3268,7 @@ static int pma_get_opa_porterrors(struct opa_pma_mad *pmp,
 	vlinfo = &rsp->vls[0];
 	vfi = 0;
 	vl_select_mask = be32_to_cpu(req->vl_select_mask);
-	for_each_set_bit(vl, (unsigned long *)&(vl_select_mask),
+	for_each_set_bit(vl, &vl_select_mask, BITS_PER_LONG) {
-			 8 * sizeof(req->vl_select_mask)) {
 		memset(vlinfo, 0, sizeof(*vlinfo));
 		rsp->vls[vfi].port_vl_xmit_discards =
 			cpu_to_be64(read_port_cntr(ppd, C_SW_XMIT_DSCD_VL,
@@ -3485,7 +3479,7 @@ static int pma_set_opa_portstatus(struct opa_pma_mad *pmp,
 	u32 nports = be32_to_cpu(pmp->mad_hdr.attr_mod) >> 24;
 	u64 portn = be64_to_cpu(req->port_select_mask[3]);
 	u32 counter_select = be32_to_cpu(req->counter_select_mask);
-	u32 vl_select_mask = VL_MASK_ALL; /* clear all per-vl cnts */
+	unsigned long vl_select_mask = VL_MASK_ALL; /* clear all per-vl cnts */
 	unsigned long vl;
 
 	if ((nports != 1) || (portn != 1 << port)) {
@@ -3579,8 +3573,7 @@ static int pma_set_opa_portstatus(struct opa_pma_mad *pmp,
 	if (counter_select & CS_UNCORRECTABLE_ERRORS)
 		write_dev_cntr(dd, C_DC_UNC_ERR, CNTR_INVALID_VL, 0);
 
-	for_each_set_bit(vl, (unsigned long *)&(vl_select_mask),
+	for_each_set_bit(vl, &vl_select_mask, BITS_PER_LONG) {
-			 8 * sizeof(vl_select_mask)) {
 		if (counter_select & CS_PORT_XMIT_DATA)
 			write_port_cntr(ppd, C_TX_FLIT_VL, idx_from_vl(vl), 0);
 

drivers/infiniband/hw/hfi1/rc.c

@@ -595,11 +595,8 @@ check_s_state:
 		case IB_WR_SEND_WITH_IMM:
 		case IB_WR_SEND_WITH_INV:
 			/* If no credit, return. */
-			if (!(qp->s_flags & RVT_S_UNLIMITED_CREDIT) &&
+			if (!rvt_rc_credit_avail(qp, wqe))
-			    rvt_cmp_msn(wqe->ssn, qp->s_lsn + 1) > 0) {
-				qp->s_flags |= RVT_S_WAIT_SSN_CREDIT;
 				goto bail;
-			}
 			if (len > pmtu) {
 				qp->s_state = OP(SEND_FIRST);
 				len = pmtu;
@@ -632,11 +629,8 @@ check_s_state:
 			goto no_flow_control;
 		case IB_WR_RDMA_WRITE_WITH_IMM:
 			/* If no credit, return. */
-			if (!(qp->s_flags & RVT_S_UNLIMITED_CREDIT) &&
+			if (!rvt_rc_credit_avail(qp, wqe))
-			    rvt_cmp_msn(wqe->ssn, qp->s_lsn + 1) > 0) {
-				qp->s_flags |= RVT_S_WAIT_SSN_CREDIT;
 				goto bail;
-			}
 no_flow_control:
 			put_ib_reth_vaddr(
 				wqe->rdma_wr.remote_addr,
@@ -1483,6 +1477,11 @@ static void update_num_rd_atomic(struct rvt_qp *qp, u32 psn,
 			req->ack_pending = cur_seg - req->comp_seg;
 			priv->pending_tid_r_segs += req->ack_pending;
 			qp->s_num_rd_atomic += req->ack_pending;
+			trace_hfi1_tid_req_update_num_rd_atomic(qp, 0,
+								wqe->wr.opcode,
+								wqe->psn,
+								wqe->lpsn,
+								req);
 		} else {
 			priv->pending_tid_r_segs += req->total_segs;
 			qp->s_num_rd_atomic += req->total_segs;

drivers/infiniband/hw/hfi1/tid_rdma.c

@@ -2646,6 +2646,9 @@ static bool handle_read_kdeth_eflags(struct hfi1_ctxtdata *rcd,
 	u32 fpsn;
 
 	lockdep_assert_held(&qp->r_lock);
+	trace_hfi1_rsp_read_kdeth_eflags(qp, ibpsn);
+	trace_hfi1_sender_read_kdeth_eflags(qp);
+	trace_hfi1_tid_read_sender_kdeth_eflags(qp, 0);
 	spin_lock(&qp->s_lock);
 	/* If the psn is out of valid range, drop the packet */
 	if (cmp_psn(ibpsn, qp->s_last_psn) < 0 ||
@@ -2710,6 +2713,8 @@ static bool handle_read_kdeth_eflags(struct hfi1_ctxtdata *rcd,
 		goto s_unlock;
 
 	req = wqe_to_tid_req(wqe);
+	trace_hfi1_tid_req_read_kdeth_eflags(qp, 0, wqe->wr.opcode, wqe->psn,
+					     wqe->lpsn, req);
 	switch (rcv_type) {
 	case RHF_RCV_TYPE_EXPECTED:
 		switch (rte) {
@@ -2724,6 +2729,9 @@ static bool handle_read_kdeth_eflags(struct hfi1_ctxtdata *rcd,
 			 * packets that could be still in the fabric.
 			 */
 			flow = &req->flows[req->clear_tail];
+			trace_hfi1_tid_flow_read_kdeth_eflags(qp,
+							      req->clear_tail,
+							      flow);
 			if (priv->s_flags & HFI1_R_TID_SW_PSN) {
 				diff = cmp_psn(psn,
 					       flow->flow_state.r_next_psn);

drivers/infiniband/hw/hfi1/trace_tid.h

@@ -627,6 +627,12 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp, index, flow)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_tid_flow_template, hfi1_tid_flow_read_kdeth_eflags,
+	TP_PROTO(struct rvt_qp *qp, int index, struct tid_rdma_flow *flow),
+	TP_ARGS(qp, index, flow)
+);
+
 DECLARE_EVENT_CLASS(/* tid_node */
 	hfi1_tid_node_template,
 	TP_PROTO(struct rvt_qp *qp, const char *msg, u32 index, u32 base,
@@ -851,6 +857,12 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp, psn)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_responder_info_template, hfi1_rsp_read_kdeth_eflags,
+	TP_PROTO(struct rvt_qp *qp, u32 psn),
+	TP_ARGS(qp, psn)
+);
+
 DECLARE_EVENT_CLASS(/* sender_info */
 	hfi1_sender_info_template,
 	TP_PROTO(struct rvt_qp *qp),
@@ -955,6 +967,12 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_sender_info_template, hfi1_sender_read_kdeth_eflags,
+	TP_PROTO(struct rvt_qp *qp),
+	TP_ARGS(qp)
+);
+
 DECLARE_EVENT_CLASS(/* tid_read_sender */
 	hfi1_tid_read_sender_template,
 	TP_PROTO(struct rvt_qp *qp, char newreq),
@@ -1015,6 +1033,12 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp, newreq)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_tid_read_sender_template, hfi1_tid_read_sender_kdeth_eflags,
+	TP_PROTO(struct rvt_qp *qp, char newreq),
+	TP_ARGS(qp, newreq)
+);
+
 DECLARE_EVENT_CLASS(/* tid_rdma_request */
 	hfi1_tid_rdma_request_template,
 	TP_PROTO(struct rvt_qp *qp, char newreq, u8 opcode, u32 psn, u32 lpsn,
@@ -1215,6 +1239,13 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp, newreq, opcode, psn, lpsn, req)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_tid_rdma_request_template, hfi1_tid_req_read_kdeth_eflags,
+	TP_PROTO(struct rvt_qp *qp, char newreq, u8 opcode, u32 psn, u32 lpsn,
+		 struct tid_rdma_request *req),
+	TP_ARGS(qp, newreq, opcode, psn, lpsn, req)
+);
+
 DEFINE_EVENT(/* event */
 	hfi1_tid_rdma_request_template, hfi1_tid_req_make_rc_ack_write,
 	TP_PROTO(struct rvt_qp *qp, char newreq, u8 opcode, u32 psn, u32 lpsn,
@@ -1229,6 +1260,13 @@ DEFINE_EVENT(/* event */
 	TP_ARGS(qp, newreq, opcode, psn, lpsn, req)
 );
 
+DEFINE_EVENT(/* event */
+	hfi1_tid_rdma_request_template, hfi1_tid_req_update_num_rd_atomic,
+	TP_PROTO(struct rvt_qp *qp, char newreq, u8 opcode, u32 psn, u32 lpsn,
+		 struct tid_rdma_request *req),
+	TP_ARGS(qp, newreq, opcode, psn, lpsn, req)
+);
+
 DECLARE_EVENT_CLASS(/* rc_rcv_err */
 	hfi1_rc_rcv_err_template,
 	TP_PROTO(struct rvt_qp *qp, u32 opcode, u32 psn, int diff),

drivers/infiniband/hw/hfi1/user_sdma.h

@@ -110,12 +110,6 @@ enum pkt_q_sdma_state {
 	SDMA_PKT_Q_DEFERRED,
 };
 
-/*
- * Maximum retry attempts to submit a TX request
- * before putting the process to sleep.
- */
-#define MAX_DEFER_RETRY_COUNT 1
-
 #define SDMA_IOWAIT_TIMEOUT 1000 /* in milliseconds */
 
 #define SDMA_DBG(req, fmt, ...)				     \

drivers/infiniband/hw/hfi1/verbs.c

@@ -874,16 +874,17 @@ int hfi1_verbs_send_dma(struct rvt_qp *qp, struct hfi1_pkt_state *ps,
 			else
 				pbc |= (ib_is_sc5(sc5) << PBC_DC_INFO_SHIFT);
 
-			if (unlikely(hfi1_dbg_should_fault_tx(qp, ps->opcode)))
-				pbc = hfi1_fault_tx(qp, ps->opcode, pbc);
 			pbc = create_pbc(ppd,
 					 pbc,
 					 qp->srate_mbps,
 					 vl,
 					 plen);
 
-			/* Update HCRC based on packet opcode */
+			if (unlikely(hfi1_dbg_should_fault_tx(qp, ps->opcode)))
-			pbc = update_hcrc(ps->opcode, pbc);
+				pbc = hfi1_fault_tx(qp, ps->opcode, pbc);
+			else
+				/* Update HCRC based on packet opcode */
+				pbc = update_hcrc(ps->opcode, pbc);
 		}
 		tx->wqe = qp->s_wqe;
 		ret = build_verbs_tx_desc(tx->sde, len, tx, ahg_info, pbc);
@@ -1030,12 +1031,12 @@ int hfi1_verbs_send_pio(struct rvt_qp *qp, struct hfi1_pkt_state *ps,
 		else
 			pbc |= (ib_is_sc5(sc5) << PBC_DC_INFO_SHIFT);
 
+		pbc = create_pbc(ppd, pbc, qp->srate_mbps, vl, plen);
 		if (unlikely(hfi1_dbg_should_fault_tx(qp, ps->opcode)))
 			pbc = hfi1_fault_tx(qp, ps->opcode, pbc);
-		pbc = create_pbc(ppd, pbc, qp->srate_mbps, vl, plen);
+		else
-
+			/* Update HCRC based on packet opcode */
-		/* Update HCRC based on packet opcode */
+			pbc = update_hcrc(ps->opcode, pbc);
-		pbc = update_hcrc(ps->opcode, pbc);
 	}
 	if (cb)
 		iowait_pio_inc(&priv->s_iowait);

drivers/infiniband/hw/hns/Kconfig

@@ -8,8 +8,6 @@ config INFINIBAND_HNS
 	  is used in Hisilicon Hip06 and more further ICT SoC based on
 	  platform device.
 
-	  To compile HIP06 or HIP08 driver as module, choose M here.
-
 config INFINIBAND_HNS_HIP06
 	tristate "Hisilicon Hip06 Family RoCE support"
 	depends on INFINIBAND_HNS && HNS && HNS_DSAF && HNS_ENET
@@ -17,15 +15,9 @@ config INFINIBAND_HNS_HIP06
 	  RoCE driver support for Hisilicon RoCE engine in Hisilicon Hip06 and
 	  Hip07 SoC. These RoCE engines are platform devices.
 
-	  To compile this driver, choose Y here: if INFINIBAND_HNS is m, this
-	  module will be called hns-roce-hw-v1
-
 config INFINIBAND_HNS_HIP08
 	tristate "Hisilicon Hip08 Family RoCE support"
 	depends on INFINIBAND_HNS && PCI && HNS3
 	---help---
 	  RoCE driver support for Hisilicon RoCE engine in Hisilicon Hip08 SoC.
 	  The RoCE engine is a PCI device.
-
-	  To compile this driver, choose Y here: if INFINIBAND_HNS is m, this
-	  module will be called hns-roce-hw-v2.

drivers/infiniband/hw/hns/hns_roce_ah.c

@@ -66,11 +66,9 @@ int hns_roce_create_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr,
 			     HNS_ROCE_VLAN_SL_SHIFT;
 	}
 
-	ah->av.port_pd = cpu_to_le32(to_hr_pd(ibah->pd)->pdn |
+	ah->av.port = rdma_ah_get_port_num(ah_attr);
-				     (rdma_ah_get_port_num(ah_attr) <<
-				     HNS_ROCE_PORT_NUM_SHIFT));
 	ah->av.gid_index = grh->sgid_index;
-	ah->av.vlan = cpu_to_le16(vlan_tag);
+	ah->av.vlan = vlan_tag;
 	ah->av.vlan_en = vlan_en;
 	dev_dbg(dev, "gid_index = 0x%x,vlan = 0x%x\n", ah->av.gid_index,
 		ah->av.vlan);
@@ -79,8 +77,7 @@ int hns_roce_create_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr,
 		ah->av.stat_rate = IB_RATE_10_GBPS;
 
 	memcpy(ah->av.dgid, grh->dgid.raw, HNS_ROCE_GID_SIZE);
-	ah->av.sl_tclass_flowlabel = cpu_to_le32(rdma_ah_get_sl(ah_attr) <<
+	ah->av.sl = rdma_ah_get_sl(ah_attr);
-						 HNS_ROCE_SL_SHIFT);
 
 	return 0;
 }
@@ -91,17 +88,11 @@ int hns_roce_query_ah(struct ib_ah *ibah, struct rdma_ah_attr *ah_attr)
 
 	memset(ah_attr, 0, sizeof(*ah_attr));
 
-	rdma_ah_set_sl(ah_attr, (le32_to_cpu(ah->av.sl_tclass_flowlabel) >>
+	rdma_ah_set_sl(ah_attr, ah->av.sl);
-				 HNS_ROCE_SL_SHIFT));
+	rdma_ah_set_port_num(ah_attr, ah->av.port);
-	rdma_ah_set_port_num(ah_attr, (le32_to_cpu(ah->av.port_pd) >>
-				       HNS_ROCE_PORT_NUM_SHIFT));
 	rdma_ah_set_static_rate(ah_attr, ah->av.stat_rate);
-	rdma_ah_set_grh(ah_attr, NULL,
+	rdma_ah_set_grh(ah_attr, NULL, ah->av.flowlabel,
-			(le32_to_cpu(ah->av.sl_tclass_flowlabel) &
+			ah->av.gid_index, ah->av.hop_limit, ah->av.tclass);
-			 HNS_ROCE_FLOW_LABEL_MASK), ah->av.gid_index,
-			ah->av.hop_limit,
-			(le32_to_cpu(ah->av.sl_tclass_flowlabel) >>
-			 HNS_ROCE_TCLASS_SHIFT));
 	rdma_ah_set_dgid_raw(ah_attr, ah->av.dgid);
 
 	return 0;

drivers/infiniband/hw/hns/hns_roce_cmd.c

@@ -211,7 +211,6 @@ int hns_roce_cmd_init(struct hns_roce_dev *hr_dev)
 	mutex_init(&hr_dev->cmd.hcr_mutex);
 	sema_init(&hr_dev->cmd.poll_sem, 1);
 	hr_dev->cmd.use_events = 0;
-	hr_dev->cmd.toggle = 1;
 	hr_dev->cmd.max_cmds = CMD_MAX_NUM;
 	hr_dev->cmd.pool = dma_pool_create("hns_roce_cmd", dev,
 					   HNS_ROCE_MAILBOX_SIZE,
@@ -252,23 +251,15 @@ int hns_roce_cmd_use_events(struct hns_roce_dev *hr_dev)
 	hr_cmd->token_mask = CMD_TOKEN_MASK;
 	hr_cmd->use_events = 1;
 
-	down(&hr_cmd->poll_sem);
-
 	return 0;
 }
 
 void hns_roce_cmd_use_polling(struct hns_roce_dev *hr_dev)
 {
 	struct hns_roce_cmdq *hr_cmd = &hr_dev->cmd;
-	int i;
-
-	hr_cmd->use_events = 0;
-
-	for (i = 0; i < hr_cmd->max_cmds; ++i)
-		down(&hr_cmd->event_sem);
 
 	kfree(hr_cmd->context);
-	up(&hr_cmd->poll_sem);
+	hr_cmd->use_events = 0;
 }
 
 struct hns_roce_cmd_mailbox

drivers/infiniband/hw/hns/hns_roce_cq.c

@@ -83,7 +83,6 @@ static int hns_roce_sw2hw_cq(struct hns_roce_dev *dev,
 
 static int hns_roce_cq_alloc(struct hns_roce_dev *hr_dev, int nent,
 			     struct hns_roce_mtt *hr_mtt,
-			     struct hns_roce_uar *hr_uar,
 			     struct hns_roce_cq *hr_cq, int vector)
 {
 	struct hns_roce_cmd_mailbox *mailbox;
@@ -154,7 +153,6 @@ static int hns_roce_cq_alloc(struct hns_roce_dev *hr_dev, int nent,
 
 	hr_cq->cons_index = 0;
 	hr_cq->arm_sn = 1;
-	hr_cq->uar = hr_uar;
 
 	atomic_set(&hr_cq->refcount, 1);
 	init_completion(&hr_cq->free);
@@ -298,21 +296,127 @@ static void hns_roce_ib_free_cq_buf(struct hns_roce_dev *hr_dev,
 			  &buf->hr_buf);
 }
 
+static int create_user_cq(struct hns_roce_dev *hr_dev,
+			  struct hns_roce_cq *hr_cq,
+			  struct ib_udata *udata,
+			  struct hns_roce_ib_create_cq_resp *resp,
+			  int cq_entries)
+{
+	struct hns_roce_ib_create_cq ucmd;
+	struct device *dev = hr_dev->dev;
+	int ret;
+	struct hns_roce_ucontext *context = rdma_udata_to_drv_context(
+				   udata, struct hns_roce_ucontext, ibucontext);
+
+	if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
+		dev_err(dev, "Failed to copy_from_udata.\n");
+		return -EFAULT;
+	}
+
+	/* Get user space address, write it into mtt table */
+	ret = hns_roce_ib_get_cq_umem(hr_dev, udata, &hr_cq->hr_buf,
+				      &hr_cq->umem, ucmd.buf_addr,
+				      cq_entries);
+	if (ret) {
+		dev_err(dev, "Failed to get_cq_umem.\n");
+		return ret;
+	}
+
+	if ((hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) &&
+	    (udata->outlen >= sizeof(*resp))) {
+		ret = hns_roce_db_map_user(context, udata, ucmd.db_addr,
+					   &hr_cq->db);
+		if (ret) {
+			dev_err(dev, "cq record doorbell map failed!\n");
+			goto err_mtt;
+		}
+		hr_cq->db_en = 1;
+		resp->cap_flags |= HNS_ROCE_SUPPORT_CQ_RECORD_DB;
+	}
+
+	return 0;
+
+err_mtt:
+	hns_roce_mtt_cleanup(hr_dev, &hr_cq->hr_buf.hr_mtt);
+	ib_umem_release(hr_cq->umem);
+
+	return ret;
+}
+
+static int create_kernel_cq(struct hns_roce_dev *hr_dev,
+			    struct hns_roce_cq *hr_cq, int cq_entries)
+{
+	struct device *dev = hr_dev->dev;
+	struct hns_roce_uar *uar;
+	int ret;
+
+	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) {
+		ret = hns_roce_alloc_db(hr_dev, &hr_cq->db, 1);
+		if (ret)
+			return ret;
+
+		hr_cq->set_ci_db = hr_cq->db.db_record;
+		*hr_cq->set_ci_db = 0;
+		hr_cq->db_en = 1;
+	}
+
+	/* Init mtt table and write buff address to mtt table */
+	ret = hns_roce_ib_alloc_cq_buf(hr_dev, &hr_cq->hr_buf, cq_entries);
+	if (ret) {
+		dev_err(dev, "Failed to alloc_cq_buf.\n");
+		goto err_db;
+	}
+
+	uar = &hr_dev->priv_uar;
+	hr_cq->cq_db_l = hr_dev->reg_base + hr_dev->odb_offset +
+			 DB_REG_OFFSET * uar->index;
+
+	return 0;
+
+err_db:
+	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB)
+		hns_roce_free_db(hr_dev, &hr_cq->db);
+
+	return ret;
+}
+
+static void destroy_user_cq(struct hns_roce_dev *hr_dev,
+			    struct hns_roce_cq *hr_cq,
+			    struct ib_udata *udata,
+			    struct hns_roce_ib_create_cq_resp *resp)
+{
+	struct hns_roce_ucontext *context = rdma_udata_to_drv_context(
+				   udata, struct hns_roce_ucontext, ibucontext);
+
+	if ((hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) &&
+	    (udata->outlen >= sizeof(*resp)))
+		hns_roce_db_unmap_user(context, &hr_cq->db);
+
+	hns_roce_mtt_cleanup(hr_dev, &hr_cq->hr_buf.hr_mtt);
+	ib_umem_release(hr_cq->umem);
+}
+
+static void destroy_kernel_cq(struct hns_roce_dev *hr_dev,
+			      struct hns_roce_cq *hr_cq)
+{
+	hns_roce_mtt_cleanup(hr_dev, &hr_cq->hr_buf.hr_mtt);
+	hns_roce_ib_free_cq_buf(hr_dev, &hr_cq->hr_buf, hr_cq->ib_cq.cqe);
+
+	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB)
+		hns_roce_free_db(hr_dev, &hr_cq->db);
+}
+
 int hns_roce_ib_create_cq(struct ib_cq *ib_cq,
 			  const struct ib_cq_init_attr *attr,
 			  struct ib_udata *udata)
 {
 	struct hns_roce_dev *hr_dev = to_hr_dev(ib_cq->device);
 	struct device *dev = hr_dev->dev;
-	struct hns_roce_ib_create_cq ucmd;
 	struct hns_roce_ib_create_cq_resp resp = {};
 	struct hns_roce_cq *hr_cq = to_hr_cq(ib_cq);
-	struct hns_roce_uar *uar = NULL;
 	int vector = attr->comp_vector;
 	int cq_entries = attr->cqe;
 	int ret;
-	struct hns_roce_ucontext *context = rdma_udata_to_drv_context(
-		udata, struct hns_roce_ucontext, ibucontext);
 
 	if (cq_entries < 1 || cq_entries > hr_dev->caps.max_cqes) {
 		dev_err(dev, "Creat CQ failed. entries=%d, max=%d\n",
@@ -328,61 +432,21 @@ int hns_roce_ib_create_cq(struct ib_cq *ib_cq,
 	spin_lock_init(&hr_cq->lock);
 
 	if (udata) {
-		if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
+		ret = create_user_cq(hr_dev, hr_cq, udata, &resp, cq_entries);
-			dev_err(dev, "Failed to copy_from_udata.\n");
-			ret = -EFAULT;
-			goto err_cq;
-		}
-
-		/* Get user space address, write it into mtt table */
-		ret = hns_roce_ib_get_cq_umem(hr_dev, udata, &hr_cq->hr_buf,
-					      &hr_cq->umem, ucmd.buf_addr,
-					      cq_entries);
 		if (ret) {
-			dev_err(dev, "Failed to get_cq_umem.\n");
+			dev_err(dev, "Create cq failed in user mode!\n");
 			goto err_cq;
 		}
-
-		if ((hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) &&
-		    (udata->outlen >= sizeof(resp))) {
-			ret = hns_roce_db_map_user(context, udata, ucmd.db_addr,
-						   &hr_cq->db);
-			if (ret) {
-				dev_err(dev, "cq record doorbell map failed!\n");
-				goto err_mtt;
-			}
-			hr_cq->db_en = 1;
-			resp.cap_flags |= HNS_ROCE_SUPPORT_CQ_RECORD_DB;
-		}
-
-		/* Get user space parameters */
-		uar = &context->uar;
 	} else {
-		if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) {
+		ret = create_kernel_cq(hr_dev, hr_cq, cq_entries);
-			ret = hns_roce_alloc_db(hr_dev, &hr_cq->db, 1);
-			if (ret)
-				goto err_cq;
-
-			hr_cq->set_ci_db = hr_cq->db.db_record;
-			*hr_cq->set_ci_db = 0;
-			hr_cq->db_en = 1;
-		}
-
-		/* Init mmt table and write buff address to mtt table */
-		ret = hns_roce_ib_alloc_cq_buf(hr_dev, &hr_cq->hr_buf,
-					       cq_entries);
 		if (ret) {
-			dev_err(dev, "Failed to alloc_cq_buf.\n");
+			dev_err(dev, "Create cq failed in kernel mode!\n");
-			goto err_db;
+			goto err_cq;
 		}
-
-		uar = &hr_dev->priv_uar;
-		hr_cq->cq_db_l = hr_dev->reg_base + hr_dev->odb_offset +
-				DB_REG_OFFSET * uar->index;
 	}
 
 	/* Allocate cq index, fill cq_context */
-	ret = hns_roce_cq_alloc(hr_dev, cq_entries, &hr_cq->hr_buf.hr_mtt, uar,
+	ret = hns_roce_cq_alloc(hr_dev, cq_entries, &hr_cq->hr_buf.hr_mtt,
 				hr_cq, vector);
 	if (ret) {
 		dev_err(dev, "Creat CQ .Failed to cq_alloc.\n");
@@ -416,20 +480,10 @@ err_cqc:
 	hns_roce_free_cq(hr_dev, hr_cq);
 
 err_dbmap:
-	if (udata && (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB) &&
+	if (udata)
-	    (udata->outlen >= sizeof(resp)))
+		destroy_user_cq(hr_dev, hr_cq, udata, &resp);
-		hns_roce_db_unmap_user(context, &hr_cq->db);
+	else
-
+		destroy_kernel_cq(hr_dev, hr_cq);
-err_mtt:
-	hns_roce_mtt_cleanup(hr_dev, &hr_cq->hr_buf.hr_mtt);
-	ib_umem_release(hr_cq->umem);
-	if (!udata)
-		hns_roce_ib_free_cq_buf(hr_dev, &hr_cq->hr_buf,
-					hr_cq->ib_cq.cqe);
-
-err_db:
-	if (!udata && (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB))
-		hns_roce_free_db(hr_dev, &hr_cq->db);
 
 err_cq:
 	return ret;

drivers/infiniband/hw/hns/hns_roce_device.h

@@ -84,7 +84,6 @@
 #define HNS_ROCE_CEQ_ENTRY_SIZE			0x4
 #define HNS_ROCE_AEQ_ENTRY_SIZE			0x10
 
-/* 4G/4K = 1M */
 #define HNS_ROCE_SL_SHIFT			28
 #define HNS_ROCE_TCLASS_SHIFT			20
 #define HNS_ROCE_FLOW_LABEL_MASK		0xfffff
@@ -128,6 +127,11 @@
 #define HNS_ROCE_IDX_QUE_ENTRY_SZ		4
 #define SRQ_DB_REG				0x230
 
+/* The chip implementation of the consumer index is calculated
+ * according to twice the actual EQ depth
+ */
+#define EQ_DEPTH_COEFF				2
+
 enum {
 	HNS_ROCE_SUPPORT_RQ_RECORD_DB = 1 << 0,
 	HNS_ROCE_SUPPORT_SQ_RECORD_DB = 1 << 1,
@@ -322,7 +326,7 @@ struct hns_roce_hem_table {
 	unsigned long	num_hem;
 	/* HEM entry record obj total num */
 	unsigned long	num_obj;
-	/*Single obj size */
+	/* Single obj size */
 	unsigned long	obj_size;
 	unsigned long	table_chunk_size;
 	int		lowmem;
@@ -343,7 +347,7 @@ struct hns_roce_mtt {
 
 struct hns_roce_buf_region {
 	int offset; /* page offset */
-	u32 count; /* page count*/
+	u32 count; /* page count */
 	int hopnum; /* addressing hop num */
 };
 
@@ -384,25 +388,25 @@ struct hns_roce_mr {
 	u64			size; /* Address range of MR */
 	u32			key; /* Key of MR */
 	u32			pd;   /* PD num of MR */
-	u32			access;/* Access permission of MR */
+	u32			access;	/* Access permission of MR */
 	u32			npages;
 	int			enabled; /* MR's active status */
 	int			type;	/* MR's register type */
-	u64			*pbl_buf;/* MR's PBL space */
+	u64			*pbl_buf;	/* MR's PBL space */
 	dma_addr_t		pbl_dma_addr;	/* MR's PBL space PA */
-	u32			pbl_size;/* PA number in the PBL */
+	u32			pbl_size;	/* PA number in the PBL */
-	u64			pbl_ba;/* page table address */
+	u64			pbl_ba;		/* page table address */
-	u32			l0_chunk_last_num;/* L0 last number */
+	u32			l0_chunk_last_num;	/* L0 last number */
-	u32			l1_chunk_last_num;/* L1 last number */
+	u32			l1_chunk_last_num;	/* L1 last number */
-	u64			**pbl_bt_l2;/* PBL BT L2 */
+	u64			**pbl_bt_l2;	/* PBL BT L2 */
-	u64			**pbl_bt_l1;/* PBL BT L1 */
+	u64			**pbl_bt_l1;	/* PBL BT L1 */
-	u64			*pbl_bt_l0;/* PBL BT L0 */
+	u64			*pbl_bt_l0;	/* PBL BT L0 */
-	dma_addr_t		*pbl_l2_dma_addr;/* PBL BT L2 dma addr */
+	dma_addr_t		*pbl_l2_dma_addr;	/* PBL BT L2 dma addr */
-	dma_addr_t		*pbl_l1_dma_addr;/* PBL BT L1 dma addr */
+	dma_addr_t		*pbl_l1_dma_addr;	/* PBL BT L1 dma addr */
-	dma_addr_t		pbl_l0_dma_addr;/* PBL BT L0 dma addr */
+	dma_addr_t		pbl_l0_dma_addr;	/* PBL BT L0 dma addr */
-	u32			pbl_ba_pg_sz;/* BT chunk page size */
+	u32			pbl_ba_pg_sz;	/* BT chunk page size */
-	u32			pbl_buf_pg_sz;/* buf chunk page size */
+	u32			pbl_buf_pg_sz;	/* buf chunk page size */
-	u32			pbl_hop_num;/* multi-hop number */
+	u32			pbl_hop_num;	/* multi-hop number */
 };
 
 struct hns_roce_mr_table {
@@ -425,16 +429,16 @@ struct hns_roce_wq {
 	u32		max_post;
 	int		max_gs;
 	int		offset;
-	int		wqe_shift;/* WQE size */
+	int		wqe_shift;	/* WQE size */
 	u32		head;
 	u32		tail;
 	void __iomem	*db_reg_l;
 };
 
 struct hns_roce_sge {
-	int		sge_cnt;  /* SGE num */
+	int		sge_cnt;	/* SGE num */
 	int		offset;
-	int		sge_shift;/* SGE size */
+	int		sge_shift;	/* SGE size */
 };
 
 struct hns_roce_buf_list {
@@ -569,14 +573,16 @@ struct hns_roce_raq_table {
 };
 
 struct hns_roce_av {
-	__le32      port_pd;
+	u8          port;
 	u8          gid_index;
 	u8          stat_rate;
 	u8          hop_limit;
-	__le32      sl_tclass_flowlabel;
+	u32         flowlabel;
+	u8          sl;
+	u8          tclass;
 	u8          dgid[HNS_ROCE_GID_SIZE];
 	u8          mac[ETH_ALEN];
-	__le16      vlan;
+	u16         vlan;
 	bool	    vlan_en;
 };
 
@@ -618,7 +624,6 @@ struct hns_roce_cmdq {
 	 * close device, switch into poll mode(non event mode)
 	 */
 	u8			use_events;
-	u8			toggle;
 };
 
 struct hns_roce_cmd_mailbox {
@@ -652,10 +657,8 @@ struct hns_roce_qp {
 	u8			rdb_en;
 	u8			sdb_en;
 	u32			doorbell_qpn;
-	__le32			sq_signal_bits;
+	u32			sq_signal_bits;
 	u32			sq_next_wqe;
-	int			sq_max_wqes_per_wr;
-	int			sq_spare_wqes;
 	struct hns_roce_wq	sq;
 
 	struct ib_umem		*umem;
@@ -709,7 +712,7 @@ enum {
 };
 
 struct hns_roce_ceqe {
-	u32			comp;
+	__le32			comp;
 };
 
 struct hns_roce_aeqe {
@@ -752,7 +755,7 @@ struct hns_roce_eq {
 	struct hns_roce_dev		*hr_dev;
 	void __iomem			*doorbell;
 
-	int				type_flag;/* Aeq:1 ceq:0 */
+	int				type_flag; /* Aeq:1 ceq:0 */
 	int				eqn;
 	u32				entries;
 	int				log_entries;
@@ -798,22 +801,22 @@ struct hns_roce_caps {
 	int		local_ca_ack_delay;
 	int		num_uars;
 	u32		phy_num_uars;
-	u32		max_sq_sg;	/* 2 */
+	u32		max_sq_sg;
-	u32		max_sq_inline;	/* 32 */
+	u32		max_sq_inline;
-	u32		max_rq_sg;	/* 2 */
+	u32		max_rq_sg;
 	u32		max_extend_sg;
-	int		num_qps;	/* 256k */
+	int		num_qps;
 	int             reserved_qps;
 	int		num_qpc_timer;
 	int		num_cqc_timer;
 	u32		max_srq_sg;
 	int		num_srqs;
-	u32		max_wqes;	/* 16k */
+	u32		max_wqes;
 	u32		max_srqs;
 	u32		max_srq_wrs;
 	u32		max_srq_sges;
-	u32		max_sq_desc_sz;	/* 64 */
+	u32		max_sq_desc_sz;
-	u32		max_rq_desc_sz;	/* 64 */
+	u32		max_rq_desc_sz;
 	u32		max_srq_desc_sz;
 	int		max_qp_init_rdma;
 	int		max_qp_dest_rdma;
@@ -824,7 +827,7 @@ struct hns_roce_caps {
 	int		reserved_cqs;
 	int		reserved_srqs;
 	u32		max_srqwqes;
-	int		num_aeq_vectors;	/* 1 */
+	int		num_aeq_vectors;
 	int		num_comp_vectors;
 	int		num_other_vectors;
 	int		num_mtpts;
@@ -905,7 +908,7 @@ struct hns_roce_caps {
 	u32		sl_num;
 	u32		tsq_buf_pg_sz;
 	u32		tpq_buf_pg_sz;
-	u32		chunk_sz;	/* chunk size in non multihop mode*/
+	u32		chunk_sz;	/* chunk size in non multihop mode */
 	u64		flags;
 };
 
@@ -991,16 +994,6 @@ struct hns_roce_hw {
 	const struct ib_device_ops *hns_roce_dev_srq_ops;
 };
 
-enum hns_phy_state {
-	HNS_ROCE_PHY_SLEEP		= 1,
-	HNS_ROCE_PHY_POLLING		= 2,
-	HNS_ROCE_PHY_DISABLED		= 3,
-	HNS_ROCE_PHY_TRAINING		= 4,
-	HNS_ROCE_PHY_LINKUP		= 5,
-	HNS_ROCE_PHY_LINKERR		= 6,
-	HNS_ROCE_PHY_TEST		= 7
-};
-
 struct hns_roce_dev {
 	struct ib_device	ib_dev;
 	struct platform_device  *pdev;
@@ -1045,8 +1038,8 @@ struct hns_roce_dev {
 	int			loop_idc;
 	u32			sdb_offset;
 	u32			odb_offset;
-	dma_addr_t		tptr_dma_addr; /*only for hw v1*/
+	dma_addr_t		tptr_dma_addr;	/* only for hw v1 */
-	u32			tptr_size; /*only for hw v1*/
+	u32			tptr_size;	/* only for hw v1 */
 	const struct hns_roce_hw *hw;
 	void			*priv;
 	struct workqueue_struct *irq_workq;

drivers/infiniband/hw/hns/hns_roce_hem.c

@@ -41,29 +41,57 @@
 
 bool hns_roce_check_whether_mhop(struct hns_roce_dev *hr_dev, u32 type)
 {
-	if ((hr_dev->caps.qpc_hop_num && type == HEM_TYPE_QPC) ||
+	int hop_num = 0;
-	    (hr_dev->caps.mpt_hop_num && type == HEM_TYPE_MTPT) ||
-	    (hr_dev->caps.cqc_hop_num && type == HEM_TYPE_CQC) ||
-	    (hr_dev->caps.srqc_hop_num && type == HEM_TYPE_SRQC) ||
-	    (hr_dev->caps.sccc_hop_num && type == HEM_TYPE_SCCC) ||
-	    (hr_dev->caps.qpc_timer_hop_num && type == HEM_TYPE_QPC_TIMER) ||
-	    (hr_dev->caps.cqc_timer_hop_num && type == HEM_TYPE_CQC_TIMER) ||
-	    (hr_dev->caps.cqe_hop_num && type == HEM_TYPE_CQE) ||
-	    (hr_dev->caps.mtt_hop_num && type == HEM_TYPE_MTT) ||
-	    (hr_dev->caps.srqwqe_hop_num && type == HEM_TYPE_SRQWQE) ||
-	    (hr_dev->caps.idx_hop_num && type == HEM_TYPE_IDX))
-		return true;
 
-	return false;
+	switch (type) {
+	case HEM_TYPE_QPC:
+		hop_num = hr_dev->caps.qpc_hop_num;
+		break;
+	case HEM_TYPE_MTPT:
+		hop_num = hr_dev->caps.mpt_hop_num;
+		break;
+	case HEM_TYPE_CQC:
+		hop_num = hr_dev->caps.cqc_hop_num;
+		break;
+	case HEM_TYPE_SRQC:
+		hop_num = hr_dev->caps.srqc_hop_num;
+		break;
+	case HEM_TYPE_SCCC:
+		hop_num = hr_dev->caps.sccc_hop_num;
+		break;
+	case HEM_TYPE_QPC_TIMER:
+		hop_num = hr_dev->caps.qpc_timer_hop_num;
+		break;
+	case HEM_TYPE_CQC_TIMER:
+		hop_num = hr_dev->caps.cqc_timer_hop_num;
+		break;
+	case HEM_TYPE_CQE:
+		hop_num = hr_dev->caps.cqe_hop_num;
+		break;
+	case HEM_TYPE_MTT:
+		hop_num = hr_dev->caps.mtt_hop_num;
+		break;
+	case HEM_TYPE_SRQWQE:
+		hop_num = hr_dev->caps.srqwqe_hop_num;
+		break;
+	case HEM_TYPE_IDX:
+		hop_num = hr_dev->caps.idx_hop_num;
+		break;
+	default:
+		return false;
+	}
+
+	return hop_num ? true : false;
 }
 
 static bool hns_roce_check_hem_null(struct hns_roce_hem **hem, u64 start_idx,
-			    u32 bt_chunk_num)
+			    u32 bt_chunk_num, u64 hem_max_num)
 {
-	int i;
+	u64 check_max_num = start_idx + bt_chunk_num;
+	u64 i;
 
-	for (i = 0; i < bt_chunk_num; i++)
+	for (i = start_idx; (i < check_max_num) && (i < hem_max_num); i++)
-		if (hem[start_idx + i])
+		if (hem[i])
 			return false;
 
 	return true;
@@ -92,17 +120,13 @@ static int hns_roce_get_bt_num(u32 table_type, u32 hop_num)
 		return 0;
 }
 
-int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
+static int get_hem_table_config(struct hns_roce_dev *hr_dev,
-			   struct hns_roce_hem_table *table, unsigned long *obj,
+				struct hns_roce_hem_mhop *mhop,
-			   struct hns_roce_hem_mhop *mhop)
+				u32 type)
 {
 	struct device *dev = hr_dev->dev;
-	u32 chunk_ba_num;
-	u32 table_idx;
-	u32 bt_num;
-	u32 chunk_size;
 
-	switch (table->type) {
+	switch (type) {
 	case HEM_TYPE_QPC:
 		mhop->buf_chunk_size = 1 << (hr_dev->caps.qpc_buf_pg_sz
 					     + PAGE_SHIFT);
@@ -193,10 +217,26 @@ int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
 		break;
 	default:
 		dev_err(dev, "Table %d not support multi-hop addressing!\n",
-			 table->type);
+			type);
 		return -EINVAL;
 	}
 
+	return 0;
+}
+
+int hns_roce_calc_hem_mhop(struct hns_roce_dev *hr_dev,
+			   struct hns_roce_hem_table *table, unsigned long *obj,
+			   struct hns_roce_hem_mhop *mhop)
+{
+	struct device *dev = hr_dev->dev;
+	u32 chunk_ba_num;
+	u32 table_idx;
+	u32 bt_num;
+	u32 chunk_size;
+
+	if (get_hem_table_config(hr_dev, mhop, table->type))
+		return -EINVAL;
+
 	if (!obj)
 		return 0;
 
@@ -324,13 +364,13 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 {
 	spinlock_t *lock = &hr_dev->bt_cmd_lock;
 	struct device *dev = hr_dev->dev;
-	unsigned long end = 0;
+	long end;
 	unsigned long flags;
 	struct hns_roce_hem_iter iter;
 	void __iomem *bt_cmd;
-	u32 bt_cmd_h_val = 0;
+	__le32 bt_cmd_val[2];
-	u32 bt_cmd_val[2];
+	__le32 bt_cmd_h = 0;
-	u32 bt_cmd_l = 0;
+	__le32 bt_cmd_l = 0;
 	u64 bt_ba = 0;
 	int ret = 0;
 
@@ -340,30 +380,20 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 
 	switch (table->type) {
 	case HEM_TYPE_QPC:
-		roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, HEM_TYPE_QPC);
-		break;
 	case HEM_TYPE_MTPT:
-		roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S,
-			       HEM_TYPE_MTPT);
-		break;
 	case HEM_TYPE_CQC:
-		roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, HEM_TYPE_CQC);
-		break;
 	case HEM_TYPE_SRQC:
-		roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
+		roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S,
+			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, table->type);
-			       HEM_TYPE_SRQC);
 		break;
 	default:
 		return ret;
 	}
-	roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_M,
+
+	roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_M,
 		       ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_S, obj);
-	roce_set_bit(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_S, 0);
+	roce_set_bit(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_S, 0);
-	roce_set_bit(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_HW_SYNS_S, 1);
+	roce_set_bit(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_HW_SYNS_S, 1);
 
 	/* Currently iter only a chunk */
 	for (hns_roce_hem_first(table->hem[i], &iter);
@@ -375,7 +405,7 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 		bt_cmd = hr_dev->reg_base + ROCEE_BT_CMD_H_REG;
 
 		end = HW_SYNC_TIMEOUT_MSECS;
-		while (end) {
+		while (end > 0) {
 			if (!(readl(bt_cmd) >> BT_CMD_SYNC_SHIFT))
 				break;
 
@@ -389,13 +419,13 @@ static int hns_roce_set_hem(struct hns_roce_dev *hr_dev,
 			return -EBUSY;
 		}
 
-		bt_cmd_l = (u32)bt_ba;
+		bt_cmd_l = cpu_to_le32(bt_ba);
-		roce_set_field(bt_cmd_h_val, ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_M,
+		roce_set_field(bt_cmd_h, ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_M,
 			       ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_S,
 			       bt_ba >> BT_BA_SHIFT);
 
 		bt_cmd_val[0] = bt_cmd_l;
-		bt_cmd_val[1] = bt_cmd_h_val;
+		bt_cmd_val[1] = bt_cmd_h;
 		hns_roce_write64_k(bt_cmd_val,
 				   hr_dev->reg_base + ROCEE_BT_CMD_L_REG);
 		spin_unlock_irqrestore(lock, flags);
@@ -457,6 +487,12 @@ static int hns_roce_table_mhop_get(struct hns_roce_dev *hr_dev,
 		return -EINVAL;
 	}
 
+	if (unlikely(hem_idx >= table->num_hem)) {
+		dev_err(dev, "Table %d exceed hem limt idx = %llu,max = %lu!\n",
+			     table->type, hem_idx, table->num_hem);
+		return -EINVAL;
+	}
+
 	mutex_lock(&table->mutex);
 
 	if (table->hem[hem_idx]) {
@@ -693,7 +729,7 @@ static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
 	if (check_whether_bt_num_2(table->type, hop_num)) {
 		start_idx = mhop.l0_idx * chunk_ba_num;
 		if (hns_roce_check_hem_null(table->hem, start_idx,
-					    chunk_ba_num)) {
+					    chunk_ba_num, table->num_hem)) {
 			if (table->type < HEM_TYPE_MTT &&
 			    hr_dev->hw->clear_hem(hr_dev, table, obj, 0))
 				dev_warn(dev, "Clear HEM base address failed.\n");
@@ -707,7 +743,7 @@ static void hns_roce_table_mhop_put(struct hns_roce_dev *hr_dev,
 		start_idx = mhop.l0_idx * chunk_ba_num * chunk_ba_num +
 			    mhop.l1_idx * chunk_ba_num;
 		if (hns_roce_check_hem_null(table->hem, start_idx,
-					    chunk_ba_num)) {
+					    chunk_ba_num, table->num_hem)) {
 			if (hr_dev->hw->clear_hem(hr_dev, table, obj, 1))
 				dev_warn(dev, "Clear HEM base address failed.\n");
 
@@ -791,7 +827,8 @@ void *hns_roce_table_find(struct hns_roce_dev *hr_dev,
 	} else {
 		u32 seg_size = 64; /* 8 bytes per BA and 8 BA per segment */
 
-		hns_roce_calc_hem_mhop(hr_dev, table, &mhop_obj, &mhop);
+		if (hns_roce_calc_hem_mhop(hr_dev, table, &mhop_obj, &mhop))
+			goto out;
 		/* mtt mhop */
 		i = mhop.l0_idx;
 		j = mhop.l1_idx;
@@ -840,11 +877,13 @@ int hns_roce_table_get_range(struct hns_roce_dev *hr_dev,
 {
 	struct hns_roce_hem_mhop mhop;
 	unsigned long inc = table->table_chunk_size / table->obj_size;
-	unsigned long i;
+	unsigned long i = 0;
 	int ret;
 
 	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
-		hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+		ret = hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+		if (ret)
+			goto fail;
 		inc = mhop.bt_chunk_size / table->obj_size;
 	}
 
@@ -874,7 +913,8 @@ void hns_roce_table_put_range(struct hns_roce_dev *hr_dev,
 	unsigned long i;
 
 	if (hns_roce_check_whether_mhop(hr_dev, table->type)) {
-		hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+		if (hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop))
+			return;
 		inc = mhop.bt_chunk_size / table->obj_size;
 	}
 
@@ -887,7 +927,6 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 			    unsigned long obj_size, unsigned long nobj,
 			    int use_lowmem)
 {
-	struct device *dev = hr_dev->dev;
 	unsigned long obj_per_chunk;
 	unsigned long num_hem;
 
@@ -900,99 +939,21 @@ int hns_roce_init_hem_table(struct hns_roce_dev *hr_dev,
 		if (!table->hem)
 			return -ENOMEM;
 	} else {
+		struct hns_roce_hem_mhop mhop = {};
 		unsigned long buf_chunk_size;
 		unsigned long bt_chunk_size;
 		unsigned long bt_chunk_num;
 		unsigned long num_bt_l0 = 0;
 		u32 hop_num;
 
-		switch (type) {
+		if (get_hem_table_config(hr_dev, &mhop, type))
-		case HEM_TYPE_QPC:
-			buf_chunk_size = 1 << (hr_dev->caps.qpc_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.qpc_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.qpc_bt_num;
-			hop_num = hr_dev->caps.qpc_hop_num;
-			break;
-		case HEM_TYPE_MTPT:
-			buf_chunk_size = 1 << (hr_dev->caps.mpt_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.mpt_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.mpt_bt_num;
-			hop_num = hr_dev->caps.mpt_hop_num;
-			break;
-		case HEM_TYPE_CQC:
-			buf_chunk_size = 1 << (hr_dev->caps.cqc_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.cqc_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.cqc_bt_num;
-			hop_num = hr_dev->caps.cqc_hop_num;
-			break;
-		case HEM_TYPE_SCCC:
-			buf_chunk_size = 1 << (hr_dev->caps.sccc_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.sccc_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.sccc_bt_num;
-			hop_num = hr_dev->caps.sccc_hop_num;
-			break;
-		case HEM_TYPE_QPC_TIMER:
-			buf_chunk_size = 1 << (hr_dev->caps.qpc_timer_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.qpc_timer_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.qpc_timer_bt_num;
-			hop_num = hr_dev->caps.qpc_timer_hop_num;
-			break;
-		case HEM_TYPE_CQC_TIMER:
-			buf_chunk_size = 1 << (hr_dev->caps.cqc_timer_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.cqc_timer_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.cqc_timer_bt_num;
-			hop_num = hr_dev->caps.cqc_timer_hop_num;
-			break;
-		case HEM_TYPE_SRQC:
-			buf_chunk_size = 1 << (hr_dev->caps.srqc_buf_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = 1 << (hr_dev->caps.srqc_ba_pg_sz
-					+ PAGE_SHIFT);
-			num_bt_l0 = hr_dev->caps.srqc_bt_num;
-			hop_num = hr_dev->caps.srqc_hop_num;
-			break;
-		case HEM_TYPE_MTT:
-			buf_chunk_size = 1 << (hr_dev->caps.mtt_ba_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = buf_chunk_size;
-			hop_num = hr_dev->caps.mtt_hop_num;
-			break;
-		case HEM_TYPE_CQE:
-			buf_chunk_size = 1 << (hr_dev->caps.cqe_ba_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = buf_chunk_size;
-			hop_num = hr_dev->caps.cqe_hop_num;
-			break;
-		case HEM_TYPE_SRQWQE:
-			buf_chunk_size = 1 << (hr_dev->caps.srqwqe_ba_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = buf_chunk_size;
-			hop_num = hr_dev->caps.srqwqe_hop_num;
-			break;
-		case HEM_TYPE_IDX:
-			buf_chunk_size = 1 << (hr_dev->caps.idx_ba_pg_sz
-					+ PAGE_SHIFT);
-			bt_chunk_size = buf_chunk_size;
-			hop_num = hr_dev->caps.idx_hop_num;
-			break;
-		default:
-			dev_err(dev,
-			  "Table %d not support to init hem table here!\n",
-			  type);
 			return -EINVAL;
-		}
+
+		buf_chunk_size = mhop.buf_chunk_size;
+		bt_chunk_size = mhop.bt_chunk_size;
+		num_bt_l0 = mhop.ba_l0_num;
+		hop_num = mhop.hop_num;
+
 		obj_per_chunk = buf_chunk_size / obj_size;
 		num_hem = (nobj + obj_per_chunk - 1) / obj_per_chunk;
 		bt_chunk_num = bt_chunk_size / BA_BYTE_LEN;
@@ -1075,7 +1036,8 @@ static void hns_roce_cleanup_mhop_hem_table(struct hns_roce_dev *hr_dev,
 	int i;
 	u64 obj;
 
-	hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop);
+	if (hns_roce_calc_hem_mhop(hr_dev, table, NULL, &mhop))
+		return;
 	buf_chunk_size = table->type < HEM_TYPE_MTT ? mhop.buf_chunk_size :
 					mhop.bt_chunk_size;
 

drivers/infiniband/hw/hns/hns_roce_hem.h

@@ -102,9 +102,9 @@ struct hns_roce_hem_mhop {
 	u32	buf_chunk_size;
 	u32	bt_chunk_size;
 	u32	ba_l0_num;
-	u32	l0_idx;/* level 0 base address table index */
+	u32	l0_idx; /* level 0 base address table index */
-	u32	l1_idx;/* level 1 base address table index */
+	u32	l1_idx; /* level 1 base address table index */
-	u32	l2_idx;/* level 2 base address table index */
+	u32	l2_idx; /* level 2 base address table index */
 };
 
 void hns_roce_free_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem *hem);

drivers/infiniband/hw/hns/hns_roce_hw_v1.c

@@ -73,7 +73,7 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
 	int ps_opcode = 0, i = 0;
 	unsigned long flags = 0;
 	void *wqe = NULL;
-	u32 doorbell[2];
+	__le32 doorbell[2];
 	int nreq = 0;
 	u32 ind = 0;
 	int ret = 0;
@@ -175,13 +175,11 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
 			roce_set_field(ud_sq_wqe->u32_36,
 				       UD_SEND_WQE_U32_36_FLOW_LABEL_M,
 				       UD_SEND_WQE_U32_36_FLOW_LABEL_S,
-				       ah->av.sl_tclass_flowlabel &
+				       ah->av.flowlabel);
-				       HNS_ROCE_FLOW_LABEL_MASK);
 			roce_set_field(ud_sq_wqe->u32_36,
 				      UD_SEND_WQE_U32_36_PRIORITY_M,
 				      UD_SEND_WQE_U32_36_PRIORITY_S,
-				      le32_to_cpu(ah->av.sl_tclass_flowlabel) >>
+				      ah->av.sl);
-				      HNS_ROCE_SL_SHIFT);
 			roce_set_field(ud_sq_wqe->u32_36,
 				       UD_SEND_WQE_U32_36_SGID_INDEX_M,
 				       UD_SEND_WQE_U32_36_SGID_INDEX_S,
@@ -195,8 +193,7 @@ static int hns_roce_v1_post_send(struct ib_qp *ibqp,
 			roce_set_field(ud_sq_wqe->u32_40,
 				       UD_SEND_WQE_U32_40_TRAFFIC_CLASS_M,
 				       UD_SEND_WQE_U32_40_TRAFFIC_CLASS_S,
-				       ah->av.sl_tclass_flowlabel >>
+				       ah->av.tclass);
-				       HNS_ROCE_TCLASS_SHIFT);
 
 			memcpy(&ud_sq_wqe->dgid[0], &ah->av.dgid[0], GID_LEN);
 
@@ -335,10 +332,10 @@ out:
 			       SQ_DOORBELL_U32_8_QPN_S, qp->doorbell_qpn);
 		roce_set_bit(sq_db.u32_8, SQ_DOORBELL_HW_SYNC_S, 1);
 
-		doorbell[0] = le32_to_cpu(sq_db.u32_4);
+		doorbell[0] = sq_db.u32_4;
-		doorbell[1] = le32_to_cpu(sq_db.u32_8);
+		doorbell[1] = sq_db.u32_8;
 
-		hns_roce_write64_k((__le32 *)doorbell, qp->sq.db_reg_l);
+		hns_roce_write64_k(doorbell, qp->sq.db_reg_l);
 		qp->sq_next_wqe = ind;
 	}
 
@@ -363,7 +360,7 @@ static int hns_roce_v1_post_recv(struct ib_qp *ibqp,
 	struct hns_roce_dev *hr_dev = to_hr_dev(ibqp->device);
 	struct device *dev = &hr_dev->pdev->dev;
 	struct hns_roce_rq_db rq_db;
-	uint32_t doorbell[2] = {0};
+	__le32 doorbell[2] = {0};
 
 	spin_lock_irqsave(&hr_qp->rq.lock, flags);
 	ind = hr_qp->rq.head & (hr_qp->rq.wqe_cnt - 1);
@@ -437,11 +434,10 @@ out:
 			roce_set_bit(rq_db.u32_8, RQ_DOORBELL_U32_8_HW_SYNC_S,
 				     1);
 
-			doorbell[0] = le32_to_cpu(rq_db.u32_4);
+			doorbell[0] = rq_db.u32_4;
-			doorbell[1] = le32_to_cpu(rq_db.u32_8);
+			doorbell[1] = rq_db.u32_8;
 
-			hns_roce_write64_k((__le32 *)doorbell,
+			hns_roce_write64_k(doorbell, hr_qp->rq.db_reg_l);
-					   hr_qp->rq.db_reg_l);
 		}
 	}
 	spin_unlock_irqrestore(&hr_qp->rq.lock, flags);
@@ -715,7 +711,7 @@ static int hns_roce_v1_rsv_lp_qp(struct hns_roce_dev *hr_dev)
 	struct ib_cq *cq;
 	struct ib_pd *pd;
 	union ib_gid dgid;
-	u64 subnet_prefix;
+	__be64 subnet_prefix;
 	int attr_mask = 0;
 	int ret;
 	int i, j;
@@ -971,7 +967,7 @@ static int hns_roce_v1_recreate_lp_qp(struct hns_roce_dev *hr_dev)
 	struct hns_roce_free_mr *free_mr;
 	struct hns_roce_v1_priv *priv;
 	struct completion comp;
-	unsigned long end = HNS_ROCE_V1_RECREATE_LP_QP_TIMEOUT_MSECS;
+	long end = HNS_ROCE_V1_RECREATE_LP_QP_TIMEOUT_MSECS;
 
 	priv = (struct hns_roce_v1_priv *)hr_dev->priv;
 	free_mr = &priv->free_mr;
@@ -991,7 +987,7 @@ static int hns_roce_v1_recreate_lp_qp(struct hns_roce_dev *hr_dev)
 
 	queue_work(free_mr->free_mr_wq, &(lp_qp_work->work));
 
-	while (end) {
+	while (end > 0) {
 		if (try_wait_for_completion(&comp))
 			return 0;
 		msleep(HNS_ROCE_V1_RECREATE_LP_QP_WAIT_VALUE);
@@ -1109,7 +1105,7 @@ static int hns_roce_v1_dereg_mr(struct hns_roce_dev *hr_dev,
 	struct hns_roce_free_mr *free_mr;
 	struct hns_roce_v1_priv *priv;
 	struct completion comp;
-	unsigned long end = HNS_ROCE_V1_FREE_MR_TIMEOUT_MSECS;
+	long end = HNS_ROCE_V1_FREE_MR_TIMEOUT_MSECS;
 	unsigned long start = jiffies;
 	int npages;
 	int ret = 0;
@@ -1139,7 +1135,7 @@ static int hns_roce_v1_dereg_mr(struct hns_roce_dev *hr_dev,
 
 	queue_work(free_mr->free_mr_wq, &(mr_work->work));
 
-	while (end) {
+	while (end > 0) {
 		if (try_wait_for_completion(&comp))
 			goto free_mr;
 		msleep(HNS_ROCE_V1_FREE_MR_WAIT_VALUE);
@@ -2165,7 +2161,7 @@ static int hns_roce_v1_req_notify_cq(struct ib_cq *ibcq,
 {
 	struct hns_roce_cq *hr_cq = to_hr_cq(ibcq);
 	u32 notification_flag;
-	__le32 doorbell[2];
+	__le32 doorbell[2] = {};
 
 	notification_flag = (flags & IB_CQ_SOLICITED_MASK) ==
 			    IB_CQ_SOLICITED ? CQ_DB_REQ_NOT : CQ_DB_REQ_NOT_SOL;
@@ -2430,7 +2426,8 @@ static int hns_roce_v1_clear_hem(struct hns_roce_dev *hr_dev,
 {
 	struct device *dev = &hr_dev->pdev->dev;
 	struct hns_roce_v1_priv *priv;
-	unsigned long end = 0, flags = 0;
+	unsigned long flags = 0;
+	long end = HW_SYNC_TIMEOUT_MSECS;
 	__le32 bt_cmd_val[2] = {0};
 	void __iomem *bt_cmd;
 	u64 bt_ba = 0;
@@ -2439,18 +2436,12 @@ static int hns_roce_v1_clear_hem(struct hns_roce_dev *hr_dev,
 
 	switch (table->type) {
 	case HEM_TYPE_QPC:
-		roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, HEM_TYPE_QPC);
 		bt_ba = priv->bt_table.qpc_buf.map >> 12;
 		break;
 	case HEM_TYPE_MTPT:
-		roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, HEM_TYPE_MTPT);
 		bt_ba = priv->bt_table.mtpt_buf.map >> 12;
 		break;
 	case HEM_TYPE_CQC:
-		roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
-			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, HEM_TYPE_CQC);
 		bt_ba = priv->bt_table.cqc_buf.map >> 12;
 		break;
 	case HEM_TYPE_SRQC:
@@ -2459,6 +2450,8 @@ static int hns_roce_v1_clear_hem(struct hns_roce_dev *hr_dev,
 	default:
 		return 0;
 	}
+	roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_M,
+			ROCEE_BT_CMD_H_ROCEE_BT_CMD_MDF_S, table->type);
 	roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_M,
 		ROCEE_BT_CMD_H_ROCEE_BT_CMD_IN_MDF_S, obj);
 	roce_set_bit(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_S, 0);
@@ -2468,7 +2461,6 @@ static int hns_roce_v1_clear_hem(struct hns_roce_dev *hr_dev,
 
 	bt_cmd = hr_dev->reg_base + ROCEE_BT_CMD_H_REG;
 
-	end = HW_SYNC_TIMEOUT_MSECS;
 	while (1) {
 		if (readl(bt_cmd) >> BT_CMD_SYNC_SHIFT) {
 			if (!end) {
@@ -2484,7 +2476,7 @@ static int hns_roce_v1_clear_hem(struct hns_roce_dev *hr_dev,
 		end -= HW_SYNC_SLEEP_TIME_INTERVAL;
 	}
 
-	bt_cmd_val[0] = (__le32)bt_ba;
+	bt_cmd_val[0] = cpu_to_le32(bt_ba);
 	roce_set_field(bt_cmd_val[1], ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_M,
 		ROCEE_BT_CMD_H_ROCEE_BT_CMD_BA_H_S, bt_ba >> 32);
 	hns_roce_write64_k(bt_cmd_val, hr_dev->reg_base + ROCEE_BT_CMD_L_REG);
@@ -2627,7 +2619,7 @@ static int hns_roce_v1_m_sqp(struct ib_qp *ibqp, const struct ib_qp_attr *attr,
 			       QP1C_BYTES_16_PORT_NUM_S, hr_qp->phy_port);
 		roce_set_bit(context->qp1c_bytes_16,
 			     QP1C_BYTES_16_SIGNALING_TYPE_S,
-			     le32_to_cpu(hr_qp->sq_signal_bits));
+			     hr_qp->sq_signal_bits);
 		roce_set_bit(context->qp1c_bytes_16, QP1C_BYTES_16_RQ_BA_FLG_S,
 			     1);
 		roce_set_bit(context->qp1c_bytes_16, QP1C_BYTES_16_SQ_BA_FLG_S,
@@ -2933,7 +2925,7 @@ static int hns_roce_v1_m_qp(struct ib_qp *ibqp, const struct ib_qp_attr *attr,
 			     1);
 		roce_set_bit(context->qpc_bytes_32,
 			     QP_CONTEXT_QPC_BYTE_32_SIGNALING_TYPE_S,
-			     le32_to_cpu(hr_qp->sq_signal_bits));
+			     hr_qp->sq_signal_bits);
 
 		port = (attr_mask & IB_QP_PORT) ? (attr->port_num - 1) :
 			hr_qp->port;
@@ -3578,7 +3570,7 @@ static int hns_roce_v1_q_qp(struct ib_qp *ibqp, struct ib_qp_attr *qp_attr,
 	qp_attr->retry_cnt = roce_get_field(context->qpc_bytes_148,
 			     QP_CONTEXT_QPC_BYTES_148_RETRY_COUNT_M,
 			     QP_CONTEXT_QPC_BYTES_148_RETRY_COUNT_S);
-	qp_attr->rnr_retry = (u8)context->rnr_retry;
+	qp_attr->rnr_retry = (u8)le32_to_cpu(context->rnr_retry);
 
 done:
 	qp_attr->cur_qp_state = qp_attr->qp_state;
@@ -4021,7 +4013,8 @@ static int hns_roce_v1_ceq_int(struct hns_roce_dev *hr_dev,
 		++eq->cons_index;
 		ceqes_found = 1;
 
-		if (eq->cons_index > 2 * hr_dev->caps.ceqe_depth - 1) {
+		if (eq->cons_index >
+		    EQ_DEPTH_COEFF * hr_dev->caps.ceqe_depth - 1) {
 			dev_warn(&eq->hr_dev->pdev->dev,
 				"cons_index overflow, set back to 0.\n");
 			eq->cons_index = 0;
@@ -4518,7 +4511,6 @@ static int hns_roce_get_cfg(struct hns_roce_dev *hr_dev)
 	struct platform_device *pdev = NULL;
 	struct net_device *netdev = NULL;
 	struct device_node *net_node;
-	struct resource *res;
 	int port_cnt = 0;
 	u8 phy_port;
 	int ret;
@@ -4557,8 +4549,7 @@ static int hns_roce_get_cfg(struct hns_roce_dev *hr_dev)
 	}
 
 	/* get the mapped register base address */
-	res = platform_get_resource(hr_dev->pdev, IORESOURCE_MEM, 0);
+	hr_dev->reg_base = devm_platform_ioremap_resource(hr_dev->pdev, 0);
-	hr_dev->reg_base = devm_ioremap_resource(dev, res);
 	if (IS_ERR(hr_dev->reg_base))
 		return PTR_ERR(hr_dev->reg_base);
 
@@ -4633,10 +4624,8 @@ static int hns_roce_get_cfg(struct hns_roce_dev *hr_dev)
 	/* fetch the interrupt numbers */
 	for (i = 0; i < HNS_ROCE_V1_MAX_IRQ_NUM; i++) {
 		hr_dev->irq[i] = platform_get_irq(hr_dev->pdev, i);
-		if (hr_dev->irq[i] <= 0) {
+		if (hr_dev->irq[i] <= 0)
-			dev_err(dev, "platform get of irq[=%d] failed!\n", i);
 			return -EINVAL;
-		}
 	}
 
 	return 0;

drivers/infiniband/hw/hns/hns_roce_hw_v2.h

@@ -96,7 +96,10 @@
 #define HNS_ROCE_V2_UC_RC_SGE_NUM_IN_WQE	2
 #define HNS_ROCE_V2_RSV_QPS			8
 
-#define HNS_ROCE_V2_HW_RST_TIMEOUT             1000
+#define HNS_ROCE_V2_HW_RST_TIMEOUT		1000
+#define HNS_ROCE_V2_HW_RST_UNINT_DELAY		100
+
+#define HNS_ROCE_V2_HW_RST_COMPLETION_WAIT	20
 
 #define HNS_ROCE_CONTEXT_HOP_NUM		1
 #define HNS_ROCE_SCCC_HOP_NUM			1
@@ -126,8 +129,6 @@
 #define HNS_ROCE_CMD_FLAG_ERR_INTR	BIT(HNS_ROCE_CMD_FLAG_ERR_INTR_SHIFT)
 
 #define HNS_ROCE_CMQ_DESC_NUM_S		3
-#define HNS_ROCE_CMQ_EN_B		16
-#define HNS_ROCE_CMQ_ENABLE		BIT(HNS_ROCE_CMQ_EN_B)
 
 #define HNS_ROCE_CMQ_SCC_CLR_DONE_CNT		5
 

drivers/infiniband/hw/hns/hns_roce_main.c

@@ -262,7 +262,8 @@ static int hns_roce_query_port(struct ib_device *ib_dev, u8 port_num,
 	props->state = (netif_running(net_dev) && netif_carrier_ok(net_dev)) ?
 			IB_PORT_ACTIVE : IB_PORT_DOWN;
 	props->phys_state = (props->state == IB_PORT_ACTIVE) ?
-			     HNS_ROCE_PHY_LINKUP : HNS_ROCE_PHY_DISABLED;
+			     IB_PORT_PHYS_STATE_LINK_UP :
+			     IB_PORT_PHYS_STATE_DISABLED;
 
 	spin_unlock_irqrestore(&hr_dev->iboe.lock, flags);
 
@@ -901,6 +902,7 @@ int hns_roce_init(struct hns_roce_dev *hr_dev)
 		goto error_failed_cmd_init;
 	}
 
+	/* EQ depends on poll mode, event mode depends on EQ */
 	ret = hr_dev->hw->init_eq(hr_dev);
 	if (ret) {
 		dev_err(dev, "eq init failed!\n");
@@ -910,8 +912,9 @@ int hns_roce_init(struct hns_roce_dev *hr_dev)
 	if (hr_dev->cmd_mod) {
 		ret = hns_roce_cmd_use_events(hr_dev);
 		if (ret) {
-			dev_err(dev, "Switch to event-driven cmd failed!\n");
+			dev_warn(dev,
-			goto error_failed_use_event;
+				 "Cmd event  mode failed, set back to poll!\n");
+			hns_roce_cmd_use_polling(hr_dev);
 		}
 	}
 
@@ -954,8 +957,6 @@ error_failed_setup_hca:
 error_failed_init_hem:
 	if (hr_dev->cmd_mod)
 		hns_roce_cmd_use_polling(hr_dev);
-
-error_failed_use_event:
 	hr_dev->hw->cleanup_eq(hr_dev);
 
 error_failed_eq_table:

drivers/infiniband/hw/hns/hns_roce_mr.c

@@ -347,49 +347,178 @@ static void hns_roce_loop_free(struct hns_roce_dev *hr_dev,
 	mr->pbl_bt_l0 = NULL;
 	mr->pbl_l0_dma_addr = 0;
 }
+static int pbl_1hop_alloc(struct hns_roce_dev *hr_dev, int npages,
+			       struct hns_roce_mr *mr, u32 pbl_bt_sz)
+{
+	struct device *dev = hr_dev->dev;
+
+	if (npages > pbl_bt_sz / 8) {
+		dev_err(dev, "npages %d is larger than buf_pg_sz!",
+			npages);
+		return -EINVAL;
+	}
+	mr->pbl_buf = dma_alloc_coherent(dev, npages * 8,
+					 &(mr->pbl_dma_addr),
+					 GFP_KERNEL);
+	if (!mr->pbl_buf)
+		return -ENOMEM;
+
+	mr->pbl_size = npages;
+	mr->pbl_ba = mr->pbl_dma_addr;
+	mr->pbl_hop_num = 1;
+	mr->pbl_ba_pg_sz = hr_dev->caps.pbl_ba_pg_sz;
+	mr->pbl_buf_pg_sz = hr_dev->caps.pbl_buf_pg_sz;
+	return 0;
+
+}
+
+
+static int pbl_2hop_alloc(struct hns_roce_dev *hr_dev, int npages,
+			       struct hns_roce_mr *mr, u32 pbl_bt_sz)
+{
+	struct device *dev = hr_dev->dev;
+	int npages_allocated;
+	u64 pbl_last_bt_num;
+	u64 pbl_bt_cnt = 0;
+	u64 size;
+	int i;
+
+	pbl_last_bt_num = (npages + pbl_bt_sz / 8 - 1) / (pbl_bt_sz / 8);
+
+	/* alloc L1 BT */
+	for (i = 0; i < pbl_bt_sz / 8; i++) {
+		if (pbl_bt_cnt + 1 < pbl_last_bt_num) {
+			size = pbl_bt_sz;
+		} else {
+			npages_allocated = i * (pbl_bt_sz / 8);
+			size = (npages - npages_allocated) * 8;
+		}
+		mr->pbl_bt_l1[i] = dma_alloc_coherent(dev, size,
+					    &(mr->pbl_l1_dma_addr[i]),
+					    GFP_KERNEL);
+		if (!mr->pbl_bt_l1[i]) {
+			hns_roce_loop_free(hr_dev, mr, 1, i, 0);
+			return -ENOMEM;
+		}
+
+		*(mr->pbl_bt_l0 + i) = mr->pbl_l1_dma_addr[i];
+
+		pbl_bt_cnt++;
+		if (pbl_bt_cnt >= pbl_last_bt_num)
+			break;
+	}
+
+	mr->l0_chunk_last_num = i + 1;
+
+	return 0;
+}
+
+static int pbl_3hop_alloc(struct hns_roce_dev *hr_dev, int npages,
+			       struct hns_roce_mr *mr, u32 pbl_bt_sz)
+{
+	struct device *dev = hr_dev->dev;
+	int mr_alloc_done = 0;
+	int npages_allocated;
+	u64 pbl_last_bt_num;
+	u64 pbl_bt_cnt = 0;
+	u64 bt_idx;
+	u64 size;
+	int i;
+	int j = 0;
+
+	pbl_last_bt_num = (npages + pbl_bt_sz / 8 - 1) / (pbl_bt_sz / 8);
+
+	mr->pbl_l2_dma_addr = kcalloc(pbl_last_bt_num,
+				      sizeof(*mr->pbl_l2_dma_addr),
+				      GFP_KERNEL);
+	if (!mr->pbl_l2_dma_addr)
+		return -ENOMEM;
+
+	mr->pbl_bt_l2 = kcalloc(pbl_last_bt_num,
+				sizeof(*mr->pbl_bt_l2),
+				GFP_KERNEL);
+	if (!mr->pbl_bt_l2)
+		goto err_kcalloc_bt_l2;
+
+	/* alloc L1, L2 BT */
+	for (i = 0; i < pbl_bt_sz / 8; i++) {
+		mr->pbl_bt_l1[i] = dma_alloc_coherent(dev, pbl_bt_sz,
+					    &(mr->pbl_l1_dma_addr[i]),
+					    GFP_KERNEL);
+		if (!mr->pbl_bt_l1[i]) {
+			hns_roce_loop_free(hr_dev, mr, 1, i, 0);
+			goto err_dma_alloc_l0;
+		}
+
+		*(mr->pbl_bt_l0 + i) = mr->pbl_l1_dma_addr[i];
+
+		for (j = 0; j < pbl_bt_sz / 8; j++) {
+			bt_idx = i * pbl_bt_sz / 8 + j;
+
+			if (pbl_bt_cnt + 1 < pbl_last_bt_num) {
+				size = pbl_bt_sz;
+			} else {
+				npages_allocated = bt_idx *
+						   (pbl_bt_sz / 8);
+				size = (npages - npages_allocated) * 8;
+			}
+			mr->pbl_bt_l2[bt_idx] = dma_alloc_coherent(
+				      dev, size,
+				      &(mr->pbl_l2_dma_addr[bt_idx]),
+				      GFP_KERNEL);
+			if (!mr->pbl_bt_l2[bt_idx]) {
+				hns_roce_loop_free(hr_dev, mr, 2, i, j);
+				goto err_dma_alloc_l0;
+			}
+
+			*(mr->pbl_bt_l1[i] + j) =
+					mr->pbl_l2_dma_addr[bt_idx];
+
+			pbl_bt_cnt++;
+			if (pbl_bt_cnt >= pbl_last_bt_num) {
+				mr_alloc_done = 1;
+				break;
+			}
+		}
+
+		if (mr_alloc_done)
+			break;
+	}
+
+	mr->l0_chunk_last_num = i + 1;
+	mr->l1_chunk_last_num = j + 1;
+
+
+	return 0;
+
+err_dma_alloc_l0:
+	kfree(mr->pbl_bt_l2);
+	mr->pbl_bt_l2 = NULL;
+
+err_kcalloc_bt_l2:
+	kfree(mr->pbl_l2_dma_addr);
+	mr->pbl_l2_dma_addr = NULL;
+
+	return -ENOMEM;
+}
+
 
 /* PBL multi hop addressing */
 static int hns_roce_mhop_alloc(struct hns_roce_dev *hr_dev, int npages,
 			       struct hns_roce_mr *mr)
 {
 	struct device *dev = hr_dev->dev;
-	int mr_alloc_done = 0;
-	int npages_allocated;
-	int i = 0, j = 0;
 	u32 pbl_bt_sz;
 	u32 mhop_num;
-	u64 pbl_last_bt_num;
-	u64 pbl_bt_cnt = 0;
-	u64 bt_idx;
-	u64 size;
 
 	mhop_num = (mr->type == MR_TYPE_FRMR ? 1 : hr_dev->caps.pbl_hop_num);
 	pbl_bt_sz = 1 << (hr_dev->caps.pbl_ba_pg_sz + PAGE_SHIFT);
-	pbl_last_bt_num = (npages + pbl_bt_sz / 8 - 1) / (pbl_bt_sz / 8);
 
 	if (mhop_num == HNS_ROCE_HOP_NUM_0)
 		return 0;
 
-	/* hop_num = 1 */
+	if (mhop_num == 1)
-	if (mhop_num == 1) {
+		return pbl_1hop_alloc(hr_dev, npages, mr, pbl_bt_sz);
-		if (npages > pbl_bt_sz / 8) {
-			dev_err(dev, "npages %d is larger than buf_pg_sz!",
-				npages);
-			return -EINVAL;
-		}
-		mr->pbl_buf = dma_alloc_coherent(dev, npages * 8,
-						 &(mr->pbl_dma_addr),
-						 GFP_KERNEL);
-		if (!mr->pbl_buf)
-			return -ENOMEM;
-
-		mr->pbl_size = npages;
-		mr->pbl_ba = mr->pbl_dma_addr;
-		mr->pbl_hop_num = mhop_num;
-		mr->pbl_ba_pg_sz = hr_dev->caps.pbl_ba_pg_sz;
-		mr->pbl_buf_pg_sz = hr_dev->caps.pbl_buf_pg_sz;
-		return 0;
-	}
 
 	mr->pbl_l1_dma_addr = kcalloc(pbl_bt_sz / 8,
 				      sizeof(*mr->pbl_l1_dma_addr),
@@ -402,100 +531,23 @@ static int hns_roce_mhop_alloc(struct hns_roce_dev *hr_dev, int npages,
 	if (!mr->pbl_bt_l1)
 		goto err_kcalloc_bt_l1;
 
-	if (mhop_num == 3) {
-		mr->pbl_l2_dma_addr = kcalloc(pbl_last_bt_num,
-					      sizeof(*mr->pbl_l2_dma_addr),
-					      GFP_KERNEL);
-		if (!mr->pbl_l2_dma_addr)
-			goto err_kcalloc_l2_dma;
-
-		mr->pbl_bt_l2 = kcalloc(pbl_last_bt_num,
-					sizeof(*mr->pbl_bt_l2),
-					GFP_KERNEL);
-		if (!mr->pbl_bt_l2)
-			goto err_kcalloc_bt_l2;
-	}
-
 	/* alloc L0 BT */
 	mr->pbl_bt_l0 = dma_alloc_coherent(dev, pbl_bt_sz,
 					   &(mr->pbl_l0_dma_addr),
 					   GFP_KERNEL);
 	if (!mr->pbl_bt_l0)
-		goto err_dma_alloc_l0;
+		goto err_kcalloc_l2_dma;
 
 	if (mhop_num == 2) {
-		/* alloc L1 BT */
+		if (pbl_2hop_alloc(hr_dev, npages, mr, pbl_bt_sz))
-		for (i = 0; i < pbl_bt_sz / 8; i++) {
+			goto err_kcalloc_l2_dma;
-			if (pbl_bt_cnt + 1 < pbl_last_bt_num) {
+	}
-				size = pbl_bt_sz;
+
-			} else {
+	if (mhop_num == 3) {
-				npages_allocated = i * (pbl_bt_sz / 8);
+		if (pbl_3hop_alloc(hr_dev, npages, mr, pbl_bt_sz))
-				size = (npages - npages_allocated) * 8;
+			goto err_kcalloc_l2_dma;
-			}
-			mr->pbl_bt_l1[i] = dma_alloc_coherent(dev, size,
-						    &(mr->pbl_l1_dma_addr[i]),
-						    GFP_KERNEL);
-			if (!mr->pbl_bt_l1[i]) {
-				hns_roce_loop_free(hr_dev, mr, 1, i, 0);
-				goto err_dma_alloc_l0;
-			}
-
-			*(mr->pbl_bt_l0 + i) = mr->pbl_l1_dma_addr[i];
-
-			pbl_bt_cnt++;
-			if (pbl_bt_cnt >= pbl_last_bt_num)
-				break;
-		}
-	} else if (mhop_num == 3) {
-		/* alloc L1, L2 BT */
-		for (i = 0; i < pbl_bt_sz / 8; i++) {
-			mr->pbl_bt_l1[i] = dma_alloc_coherent(dev, pbl_bt_sz,
-						    &(mr->pbl_l1_dma_addr[i]),
-						    GFP_KERNEL);
-			if (!mr->pbl_bt_l1[i]) {
-				hns_roce_loop_free(hr_dev, mr, 1, i, 0);
-				goto err_dma_alloc_l0;
-			}
-
-			*(mr->pbl_bt_l0 + i) = mr->pbl_l1_dma_addr[i];
-
-			for (j = 0; j < pbl_bt_sz / 8; j++) {
-				bt_idx = i * pbl_bt_sz / 8 + j;
-
-				if (pbl_bt_cnt + 1 < pbl_last_bt_num) {
-					size = pbl_bt_sz;
-				} else {
-					npages_allocated = bt_idx *
-							   (pbl_bt_sz / 8);
-					size = (npages - npages_allocated) * 8;
-				}
-				mr->pbl_bt_l2[bt_idx] = dma_alloc_coherent(
-					      dev, size,
-					      &(mr->pbl_l2_dma_addr[bt_idx]),
-					      GFP_KERNEL);
-				if (!mr->pbl_bt_l2[bt_idx]) {
-					hns_roce_loop_free(hr_dev, mr, 2, i, j);
-					goto err_dma_alloc_l0;
-				}
-
-				*(mr->pbl_bt_l1[i] + j) =
-						mr->pbl_l2_dma_addr[bt_idx];
-
-				pbl_bt_cnt++;
-				if (pbl_bt_cnt >= pbl_last_bt_num) {
-					mr_alloc_done = 1;
-					break;
-				}
-			}
-
-			if (mr_alloc_done)
-				break;
-		}
 	}
 
-	mr->l0_chunk_last_num = i + 1;
-	if (mhop_num == 3)
-		mr->l1_chunk_last_num = j + 1;
 
 	mr->pbl_size = npages;
 	mr->pbl_ba = mr->pbl_l0_dma_addr;
@@ -505,14 +557,6 @@ static int hns_roce_mhop_alloc(struct hns_roce_dev *hr_dev, int npages,
 
 	return 0;
 
-err_dma_alloc_l0:
-	kfree(mr->pbl_bt_l2);
-	mr->pbl_bt_l2 = NULL;
-
-err_kcalloc_bt_l2:
-	kfree(mr->pbl_l2_dma_addr);
-	mr->pbl_l2_dma_addr = NULL;
-
 err_kcalloc_l2_dma:
 	kfree(mr->pbl_bt_l1);
 	mr->pbl_bt_l1 = NULL;
@@ -1161,6 +1205,83 @@ err_free:
 	return ERR_PTR(ret);
 }
 
+static int rereg_mr_trans(struct ib_mr *ibmr, int flags,
+			  u64 start, u64 length,
+			  u64 virt_addr, int mr_access_flags,
+			  struct hns_roce_cmd_mailbox *mailbox,
+			  u32 pdn, struct ib_udata *udata)
+{
+	struct hns_roce_dev *hr_dev = to_hr_dev(ibmr->device);
+	struct hns_roce_mr *mr = to_hr_mr(ibmr);
+	struct device *dev = hr_dev->dev;
+	int npages;
+	int ret;
+
+	if (mr->size != ~0ULL) {
+		npages = ib_umem_page_count(mr->umem);
+
+		if (hr_dev->caps.pbl_hop_num)
+			hns_roce_mhop_free(hr_dev, mr);
+		else
+			dma_free_coherent(dev, npages * 8,
+					  mr->pbl_buf, mr->pbl_dma_addr);
+	}
+	ib_umem_release(mr->umem);
+
+	mr->umem = ib_umem_get(udata, start, length, mr_access_flags, 0);
+	if (IS_ERR(mr->umem)) {
+		ret = PTR_ERR(mr->umem);
+		mr->umem = NULL;
+		return -ENOMEM;
+	}
+	npages = ib_umem_page_count(mr->umem);
+
+	if (hr_dev->caps.pbl_hop_num) {
+		ret = hns_roce_mhop_alloc(hr_dev, npages, mr);
+		if (ret)
+			goto release_umem;
+	} else {
+		mr->pbl_buf = dma_alloc_coherent(dev, npages * 8,
+						 &(mr->pbl_dma_addr),
+						 GFP_KERNEL);
+		if (!mr->pbl_buf) {
+			ret = -ENOMEM;
+			goto release_umem;
+		}
+	}
+
+	ret = hr_dev->hw->rereg_write_mtpt(hr_dev, mr, flags, pdn,
+					   mr_access_flags, virt_addr,
+					   length, mailbox->buf);
+	if (ret)
+		goto release_umem;
+
+
+	ret = hns_roce_ib_umem_write_mr(hr_dev, mr, mr->umem);
+	if (ret) {
+		if (mr->size != ~0ULL) {
+			npages = ib_umem_page_count(mr->umem);
+
+			if (hr_dev->caps.pbl_hop_num)
+				hns_roce_mhop_free(hr_dev, mr);
+			else
+				dma_free_coherent(dev, npages * 8,
+						  mr->pbl_buf,
+						  mr->pbl_dma_addr);
+		}
+
+		goto release_umem;
+	}
+
+	return 0;
+
+release_umem:
+	ib_umem_release(mr->umem);
+	return ret;
+
+}
+
+
 int hns_roce_rereg_user_mr(struct ib_mr *ibmr, int flags, u64 start, u64 length,
 			   u64 virt_addr, int mr_access_flags, struct ib_pd *pd,
 			   struct ib_udata *udata)
@@ -1171,7 +1292,6 @@ int hns_roce_rereg_user_mr(struct ib_mr *ibmr, int flags, u64 start, u64 length,
 	struct device *dev = hr_dev->dev;
 	unsigned long mtpt_idx;
 	u32 pdn = 0;
-	int npages;
 	int ret;
 
 	if (!mr->enabled)
@@ -1198,73 +1318,25 @@ int hns_roce_rereg_user_mr(struct ib_mr *ibmr, int flags, u64 start, u64 length,
 		pdn = to_hr_pd(pd)->pdn;
 
 	if (flags & IB_MR_REREG_TRANS) {
-		if (mr->size != ~0ULL) {
+		ret = rereg_mr_trans(ibmr, flags,
-			npages = ib_umem_page_count(mr->umem);
+				     start, length,
-
+				     virt_addr, mr_access_flags,
-			if (hr_dev->caps.pbl_hop_num)
+				     mailbox, pdn, udata);
-				hns_roce_mhop_free(hr_dev, mr);
+		if (ret)
-			else
-				dma_free_coherent(dev, npages * 8, mr->pbl_buf,
-						  mr->pbl_dma_addr);
-		}
-		ib_umem_release(mr->umem);
-
-		mr->umem =
-			ib_umem_get(udata, start, length, mr_access_flags, 0);
-		if (IS_ERR(mr->umem)) {
-			ret = PTR_ERR(mr->umem);
-			mr->umem = NULL;
 			goto free_cmd_mbox;
-		}
+	} else {
-		npages = ib_umem_page_count(mr->umem);
+		ret = hr_dev->hw->rereg_write_mtpt(hr_dev, mr, flags, pdn,
-
+						   mr_access_flags, virt_addr,
-		if (hr_dev->caps.pbl_hop_num) {
+						   length, mailbox->buf);
-			ret = hns_roce_mhop_alloc(hr_dev, npages, mr);
+		if (ret)
-			if (ret)
-				goto release_umem;
-		} else {
-			mr->pbl_buf = dma_alloc_coherent(dev, npages * 8,
-							 &(mr->pbl_dma_addr),
-							 GFP_KERNEL);
-			if (!mr->pbl_buf) {
-				ret = -ENOMEM;
-				goto release_umem;
-			}
-		}
-	}
-
-	ret = hr_dev->hw->rereg_write_mtpt(hr_dev, mr, flags, pdn,
-					   mr_access_flags, virt_addr,
-					   length, mailbox->buf);
-	if (ret) {
-		if (flags & IB_MR_REREG_TRANS)
-			goto release_umem;
-		else
 			goto free_cmd_mbox;
 	}
 
-	if (flags & IB_MR_REREG_TRANS) {
-		ret = hns_roce_ib_umem_write_mr(hr_dev, mr, mr->umem);
-		if (ret) {
-			if (mr->size != ~0ULL) {
-				npages = ib_umem_page_count(mr->umem);
-
-				if (hr_dev->caps.pbl_hop_num)
-					hns_roce_mhop_free(hr_dev, mr);
-				else
-					dma_free_coherent(dev, npages * 8,
-							  mr->pbl_buf,
-							  mr->pbl_dma_addr);
-			}
-
-			goto release_umem;
-		}
-	}
-
 	ret = hns_roce_sw2hw_mpt(hr_dev, mailbox, mtpt_idx);
 	if (ret) {
 		dev_err(dev, "SW2HW_MPT failed (%d)\n", ret);
-		goto release_umem;
+		ib_umem_release(mr->umem);
+		goto free_cmd_mbox;
 	}
 
 	mr->enabled = 1;
@@ -1275,9 +1347,6 @@ int hns_roce_rereg_user_mr(struct ib_mr *ibmr, int flags, u64 start, u64 length,
 
 	return 0;
 
-release_umem:
-	ib_umem_release(mr->umem);
-
 free_cmd_mbox:
 	hns_roce_free_cmd_mailbox(hr_dev, mailbox);
 
@@ -1357,7 +1426,7 @@ static int hns_roce_set_page(struct ib_mr *ibmr, u64 addr)
 {
 	struct hns_roce_mr *mr = to_hr_mr(ibmr);
 
-	mr->pbl_buf[mr->npages++] = cpu_to_le64(addr);
+	mr->pbl_buf[mr->npages++] = addr;
 
 	return 0;
 }
@@ -1528,10 +1597,9 @@ static int hns_roce_write_mtr(struct hns_roce_dev *hr_dev,
 		/* Save page addr, low 12 bits : 0 */
 		for (i = 0; i < count; i++) {
 			if (hr_dev->hw_rev == HNS_ROCE_HW_VER1)
-				mtts[i] = cpu_to_le64(bufs[npage] >>
+				mtts[i] = bufs[npage] >> PAGE_ADDR_SHIFT;
-							PAGE_ADDR_SHIFT);
 			else
-				mtts[i] = cpu_to_le64(bufs[npage]);
+				mtts[i] = bufs[npage];
 
 			npage++;
 		}

drivers/infiniband/hw/hns/hns_roce_qp.c

@@ -324,31 +324,46 @@ static int hns_roce_set_rq_size(struct hns_roce_dev *hr_dev,
 	return 0;
 }
 
-static int hns_roce_set_user_sq_size(struct hns_roce_dev *hr_dev,
+static int check_sq_size_with_integrity(struct hns_roce_dev *hr_dev,
-				     struct ib_qp_cap *cap,
+					struct ib_qp_cap *cap,
-				     struct hns_roce_qp *hr_qp,
+					struct hns_roce_ib_create_qp *ucmd)
-				     struct hns_roce_ib_create_qp *ucmd)
 {
 	u32 roundup_sq_stride = roundup_pow_of_two(hr_dev->caps.max_sq_desc_sz);
 	u8 max_sq_stride = ilog2(roundup_sq_stride);
-	u32 ex_sge_num;
-	u32 page_size;
-	u32 max_cnt;
 
 	/* Sanity check SQ size before proceeding */
 	if ((u32)(1 << ucmd->log_sq_bb_count) > hr_dev->caps.max_wqes ||
 	     ucmd->log_sq_stride > max_sq_stride ||
 	     ucmd->log_sq_stride < HNS_ROCE_IB_MIN_SQ_STRIDE) {
-		dev_err(hr_dev->dev, "check SQ size error!\n");
+		ibdev_err(&hr_dev->ib_dev, "check SQ size error!\n");
 		return -EINVAL;
 	}
 
 	if (cap->max_send_sge > hr_dev->caps.max_sq_sg) {
-		dev_err(hr_dev->dev, "SQ sge error! max_send_sge=%d\n",
+		ibdev_err(&hr_dev->ib_dev, "SQ sge error! max_send_sge=%d\n",
-			cap->max_send_sge);
+			  cap->max_send_sge);
 		return -EINVAL;
 	}
 
+	return 0;
+}
+
+static int hns_roce_set_user_sq_size(struct hns_roce_dev *hr_dev,
+				     struct ib_qp_cap *cap,
+				     struct hns_roce_qp *hr_qp,
+				     struct hns_roce_ib_create_qp *ucmd)
+{
+	u32 ex_sge_num;
+	u32 page_size;
+	u32 max_cnt;
+	int ret;
+
+	ret = check_sq_size_with_integrity(hr_dev, cap, ucmd);
+	if (ret) {
+		ibdev_err(&hr_dev->ib_dev, "Sanity check sq size failed\n");
+		return ret;
+	}
+
 	hr_qp->sq.wqe_cnt = 1 << ucmd->log_sq_bb_count;
 	hr_qp->sq.wqe_shift = ucmd->log_sq_stride;
 
@@ -501,43 +516,10 @@ static int calc_wqe_bt_page_shift(struct hns_roce_dev *hr_dev,
 	return bt_pg_shift - PAGE_SHIFT;
 }
 
-static int hns_roce_set_kernel_sq_size(struct hns_roce_dev *hr_dev,
+static int set_extend_sge_param(struct hns_roce_dev *hr_dev,
-				       struct ib_qp_cap *cap,
+				struct hns_roce_qp *hr_qp)
-				       struct hns_roce_qp *hr_qp)
 {
 	struct device *dev = hr_dev->dev;
-	u32 page_size;
-	u32 max_cnt;
-	int size;
-
-	if (cap->max_send_wr  > hr_dev->caps.max_wqes  ||
-	    cap->max_send_sge > hr_dev->caps.max_sq_sg ||
-	    cap->max_inline_data > hr_dev->caps.max_sq_inline) {
-		dev_err(dev, "SQ WR or sge or inline data error!\n");
-		return -EINVAL;
-	}
-
-	hr_qp->sq.wqe_shift = ilog2(hr_dev->caps.max_sq_desc_sz);
-	hr_qp->sq_max_wqes_per_wr = 1;
-	hr_qp->sq_spare_wqes = 0;
-
-	if (hr_dev->caps.min_wqes)
-		max_cnt = max(cap->max_send_wr, hr_dev->caps.min_wqes);
-	else
-		max_cnt = cap->max_send_wr;
-
-	hr_qp->sq.wqe_cnt = roundup_pow_of_two(max_cnt);
-	if ((u32)hr_qp->sq.wqe_cnt > hr_dev->caps.max_wqes) {
-		dev_err(dev, "while setting kernel sq size, sq.wqe_cnt too large\n");
-		return -EINVAL;
-	}
-
-	/* Get data_seg numbers */
-	max_cnt = max(1U, cap->max_send_sge);
-	if (hr_dev->caps.max_sq_sg <= 2)
-		hr_qp->sq.max_gs = roundup_pow_of_two(max_cnt);
-	else
-		hr_qp->sq.max_gs = max_cnt;
 
 	if (hr_qp->sq.max_gs > 2) {
 		hr_qp->sge.sge_cnt = roundup_pow_of_two(hr_qp->sq.wqe_cnt *
@@ -560,6 +542,52 @@ static int hns_roce_set_kernel_sq_size(struct hns_roce_dev *hr_dev,
 		}
 	}
 
+	return 0;
+}
+
+static int hns_roce_set_kernel_sq_size(struct hns_roce_dev *hr_dev,
+				       struct ib_qp_cap *cap,
+				       struct hns_roce_qp *hr_qp)
+{
+	struct device *dev = hr_dev->dev;
+	u32 page_size;
+	u32 max_cnt;
+	int size;
+	int ret;
+
+	if (cap->max_send_wr  > hr_dev->caps.max_wqes  ||
+	    cap->max_send_sge > hr_dev->caps.max_sq_sg ||
+	    cap->max_inline_data > hr_dev->caps.max_sq_inline) {
+		dev_err(dev, "SQ WR or sge or inline data error!\n");
+		return -EINVAL;
+	}
+
+	hr_qp->sq.wqe_shift = ilog2(hr_dev->caps.max_sq_desc_sz);
+
+	if (hr_dev->caps.min_wqes)
+		max_cnt = max(cap->max_send_wr, hr_dev->caps.min_wqes);
+	else
+		max_cnt = cap->max_send_wr;
+
+	hr_qp->sq.wqe_cnt = roundup_pow_of_two(max_cnt);
+	if ((u32)hr_qp->sq.wqe_cnt > hr_dev->caps.max_wqes) {
+		dev_err(dev, "while setting kernel sq size, sq.wqe_cnt too large\n");
+		return -EINVAL;
+	}
+
+	/* Get data_seg numbers */
+	max_cnt = max(1U, cap->max_send_sge);
+	if (hr_dev->caps.max_sq_sg <= 2)
+		hr_qp->sq.max_gs = roundup_pow_of_two(max_cnt);
+	else
+		hr_qp->sq.max_gs = max_cnt;
+
+	ret = set_extend_sge_param(hr_dev, hr_qp);
+	if (ret) {
+		dev_err(dev, "set extend sge parameters fail\n");
+		return ret;
+	}
+
 	/* Get buf size, SQ and RQ are aligned to PAGE_SIZE */
 	page_size = 1 << (hr_dev->caps.mtt_buf_pg_sz + PAGE_SHIFT);
 	hr_qp->sq.offset = 0;
@@ -607,13 +635,57 @@ static int hns_roce_qp_has_rq(struct ib_qp_init_attr *attr)
 	return 1;
 }
 
+static int alloc_rq_inline_buf(struct hns_roce_qp *hr_qp,
+			       struct ib_qp_init_attr *init_attr)
+{
+	u32 max_recv_sge = init_attr->cap.max_recv_sge;
+	struct hns_roce_rinl_wqe *wqe_list;
+	u32 wqe_cnt = hr_qp->rq.wqe_cnt;
+	int i;
+
+	/* allocate recv inline buf */
+	wqe_list = kcalloc(wqe_cnt, sizeof(struct hns_roce_rinl_wqe),
+			   GFP_KERNEL);
+
+	if (!wqe_list)
+		goto err;
+
+	/* Allocate a continuous buffer for all inline sge we need */
+	wqe_list[0].sg_list = kcalloc(wqe_cnt, (max_recv_sge *
+				      sizeof(struct hns_roce_rinl_sge)),
+				      GFP_KERNEL);
+	if (!wqe_list[0].sg_list)
+		goto err_wqe_list;
+
+	/* Assign buffers of sg_list to each inline wqe */
+	for (i = 1; i < wqe_cnt; i++)
+		wqe_list[i].sg_list = &wqe_list[0].sg_list[i * max_recv_sge];
+
+	hr_qp->rq_inl_buf.wqe_list = wqe_list;
+	hr_qp->rq_inl_buf.wqe_cnt = wqe_cnt;
+
+	return 0;
+
+err_wqe_list:
+	kfree(wqe_list);
+
+err:
+	return -ENOMEM;
+}
+
+static void free_rq_inline_buf(struct hns_roce_qp *hr_qp)
+{
+	kfree(hr_qp->rq_inl_buf.wqe_list[0].sg_list);
+	kfree(hr_qp->rq_inl_buf.wqe_list);
+}
+
 static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 				     struct ib_pd *ib_pd,
 				     struct ib_qp_init_attr *init_attr,
 				     struct ib_udata *udata, unsigned long sqpn,
 				     struct hns_roce_qp *hr_qp)
 {
-	dma_addr_t *buf_list[ARRAY_SIZE(hr_qp->regions)] = { 0 };
+	dma_addr_t *buf_list[ARRAY_SIZE(hr_qp->regions)] = { NULL };
 	struct device *dev = hr_dev->dev;
 	struct hns_roce_ib_create_qp ucmd;
 	struct hns_roce_ib_create_qp_resp resp = {};
@@ -635,9 +707,9 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 	hr_qp->ibqp.qp_type = init_attr->qp_type;
 
 	if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
-		hr_qp->sq_signal_bits = cpu_to_le32(IB_SIGNAL_ALL_WR);
+		hr_qp->sq_signal_bits = IB_SIGNAL_ALL_WR;
 	else
-		hr_qp->sq_signal_bits = cpu_to_le32(IB_SIGNAL_REQ_WR);
+		hr_qp->sq_signal_bits = IB_SIGNAL_REQ_WR;
 
 	ret = hns_roce_set_rq_size(hr_dev, &init_attr->cap, udata,
 				   hns_roce_qp_has_rq(init_attr), hr_qp);
@@ -648,33 +720,11 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 
 	if ((hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RQ_INLINE) &&
 	    hns_roce_qp_has_rq(init_attr)) {
-		/* allocate recv inline buf */
+		ret = alloc_rq_inline_buf(hr_qp, init_attr);
-		hr_qp->rq_inl_buf.wqe_list = kcalloc(hr_qp->rq.wqe_cnt,
+		if (ret) {
-					       sizeof(struct hns_roce_rinl_wqe),
+			dev_err(dev, "allocate receive inline buffer failed\n");
-					       GFP_KERNEL);
-		if (!hr_qp->rq_inl_buf.wqe_list) {
-			ret = -ENOMEM;
 			goto err_out;
 		}
-
-		hr_qp->rq_inl_buf.wqe_cnt = hr_qp->rq.wqe_cnt;
-
-		/* Firstly, allocate a list of sge space buffer */
-		hr_qp->rq_inl_buf.wqe_list[0].sg_list =
-					kcalloc(hr_qp->rq_inl_buf.wqe_cnt,
-					       init_attr->cap.max_recv_sge *
-					       sizeof(struct hns_roce_rinl_sge),
-					       GFP_KERNEL);
-		if (!hr_qp->rq_inl_buf.wqe_list[0].sg_list) {
-			ret = -ENOMEM;
-			goto err_wqe_list;
-		}
-
-		for (i = 1; i < hr_qp->rq_inl_buf.wqe_cnt; i++)
-			/* Secondly, reallocate the buffer */
-			hr_qp->rq_inl_buf.wqe_list[i].sg_list =
-				&hr_qp->rq_inl_buf.wqe_list[0].sg_list[i *
-				init_attr->cap.max_recv_sge];
 	}
 
 	page_shift = PAGE_SHIFT + hr_dev->caps.mtt_buf_pg_sz;
@@ -682,14 +732,14 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 		if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd))) {
 			dev_err(dev, "ib_copy_from_udata error for create qp\n");
 			ret = -EFAULT;
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 
 		ret = hns_roce_set_user_sq_size(hr_dev, &init_attr->cap, hr_qp,
 						&ucmd);
 		if (ret) {
 			dev_err(dev, "hns_roce_set_user_sq_size error for create qp\n");
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 
 		hr_qp->umem = ib_umem_get(udata, ucmd.buf_addr,
@@ -697,7 +747,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 		if (IS_ERR(hr_qp->umem)) {
 			dev_err(dev, "ib_umem_get error for create qp\n");
 			ret = PTR_ERR(hr_qp->umem);
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 		hr_qp->region_cnt = split_wqe_buf_region(hr_dev, hr_qp,
 				hr_qp->regions, ARRAY_SIZE(hr_qp->regions),
@@ -758,13 +808,13 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 		    IB_QP_CREATE_BLOCK_MULTICAST_LOOPBACK) {
 			dev_err(dev, "init_attr->create_flags error!\n");
 			ret = -EINVAL;
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 
 		if (init_attr->create_flags & IB_QP_CREATE_IPOIB_UD_LSO) {
 			dev_err(dev, "init_attr->create_flags error!\n");
 			ret = -EINVAL;
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 
 		/* Set SQ size */
@@ -772,7 +822,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 						  hr_qp);
 		if (ret) {
 			dev_err(dev, "hns_roce_set_kernel_sq_size error!\n");
-			goto err_rq_sge_list;
+			goto err_alloc_rq_inline_buf;
 		}
 
 		/* QP doorbell register address */
@@ -786,7 +836,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 			ret = hns_roce_alloc_db(hr_dev, &hr_qp->rdb, 0);
 			if (ret) {
 				dev_err(dev, "rq record doorbell alloc failed!\n");
-				goto err_rq_sge_list;
+				goto err_alloc_rq_inline_buf;
 			}
 			*hr_qp->rdb.db_record = 0;
 			hr_qp->rdb_en = 1;
@@ -826,11 +876,18 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 
 		hr_qp->sq.wrid = kcalloc(hr_qp->sq.wqe_cnt, sizeof(u64),
 					 GFP_KERNEL);
-		hr_qp->rq.wrid = kcalloc(hr_qp->rq.wqe_cnt, sizeof(u64),
+		if (ZERO_OR_NULL_PTR(hr_qp->sq.wrid)) {
-					 GFP_KERNEL);
-		if (!hr_qp->sq.wrid || !hr_qp->rq.wrid) {
 			ret = -ENOMEM;
-			goto err_wrid;
+			goto err_get_bufs;
+		}
+
+		if (hr_qp->rq.wqe_cnt) {
+			hr_qp->rq.wrid = kcalloc(hr_qp->rq.wqe_cnt, sizeof(u64),
+						 GFP_KERNEL);
+			if (ZERO_OR_NULL_PTR(hr_qp->rq.wrid)) {
+				ret = -ENOMEM;
+				goto err_sq_wrid;
+			}
 		}
 	}
 
@@ -875,7 +932,7 @@ static int hns_roce_create_qp_common(struct hns_roce_dev *hr_dev,
 	if (sqpn)
 		hr_qp->doorbell_qpn = 1;
 	else
-		hr_qp->doorbell_qpn = cpu_to_le64(hr_qp->qpn);
+		hr_qp->doorbell_qpn = (u32)hr_qp->qpn;
 
 	if (udata) {
 		ret = ib_copy_to_udata(udata, &resp,
@@ -916,8 +973,8 @@ err_wrid:
 		    hns_roce_qp_has_rq(init_attr))
 			hns_roce_db_unmap_user(uctx, &hr_qp->rdb);
 	} else {
-		kfree(hr_qp->sq.wrid);
+		if (hr_qp->rq.wqe_cnt)
-		kfree(hr_qp->rq.wrid);
+			kfree(hr_qp->rq.wrid);
 	}
 
 err_sq_dbmap:
@@ -928,6 +985,10 @@ err_sq_dbmap:
 		    hns_roce_qp_has_sq(init_attr))
 			hns_roce_db_unmap_user(uctx, &hr_qp->sdb);
 
+err_sq_wrid:
+	if (!udata)
+		kfree(hr_qp->sq.wrid);
+
 err_get_bufs:
 	hns_roce_free_buf_list(buf_list, hr_qp->region_cnt);
 
@@ -941,13 +1002,10 @@ err_db:
 	    (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RECORD_DB))
 		hns_roce_free_db(hr_dev, &hr_qp->rdb);
 
-err_rq_sge_list:
+err_alloc_rq_inline_buf:
-	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RQ_INLINE)
+	if ((hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RQ_INLINE) &&
-		kfree(hr_qp->rq_inl_buf.wqe_list[0].sg_list);
+	     hns_roce_qp_has_rq(init_attr))
-
+		free_rq_inline_buf(hr_qp);
-err_wqe_list:
-	if (hr_dev->caps.flags & HNS_ROCE_CAP_FLAG_RQ_INLINE)
-		kfree(hr_qp->rq_inl_buf.wqe_list);
 
 err_out:
 	return ret;
@@ -958,7 +1016,7 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *pd,
 				 struct ib_udata *udata)
 {
 	struct hns_roce_dev *hr_dev = to_hr_dev(pd->device);
-	struct device *dev = hr_dev->dev;
+	struct ib_device *ibdev = &hr_dev->ib_dev;
 	struct hns_roce_sqp *hr_sqp;
 	struct hns_roce_qp *hr_qp;
 	int ret;
@@ -972,7 +1030,8 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *pd,
 		ret = hns_roce_create_qp_common(hr_dev, pd, init_attr, udata, 0,
 						hr_qp);
 		if (ret) {
-			dev_err(dev, "Create RC QP failed\n");
+			ibdev_err(ibdev, "Create RC QP 0x%06lx failed(%d)\n",
+				  hr_qp->qpn, ret);
 			kfree(hr_qp);
 			return ERR_PTR(ret);
 		}
@@ -984,7 +1043,7 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *pd,
 	case IB_QPT_GSI: {
 		/* Userspace is not allowed to create special QPs: */
 		if (udata) {
-			dev_err(dev, "not support usr space GSI\n");
+			ibdev_err(ibdev, "not support usr space GSI\n");
 			return ERR_PTR(-EINVAL);
 		}
 
@@ -1006,7 +1065,7 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *pd,
 		ret = hns_roce_create_qp_common(hr_dev, pd, init_attr, udata,
 						hr_qp->ibqp.qp_num, hr_qp);
 		if (ret) {
-			dev_err(dev, "Create GSI QP failed!\n");
+			ibdev_err(ibdev, "Create GSI QP failed!\n");
 			kfree(hr_sqp);
 			return ERR_PTR(ret);
 		}
@@ -1014,7 +1073,8 @@ struct ib_qp *hns_roce_create_qp(struct ib_pd *pd,
 		break;
 	}
 	default:{
-		dev_err(dev, "not support QP type %d\n", init_attr->qp_type);
+		ibdev_err(ibdev, "not support QP type %d\n",
+			  init_attr->qp_type);
 		return ERR_PTR(-EINVAL);
 	}
 	}
@@ -1040,23 +1100,88 @@ int to_hr_qp_type(int qp_type)
 	return transport_type;
 }
 
+static int check_mtu_validate(struct hns_roce_dev *hr_dev,
+			      struct hns_roce_qp *hr_qp,
+			      struct ib_qp_attr *attr, int attr_mask)
+{
+	enum ib_mtu active_mtu;
+	int p;
+
+	p = attr_mask & IB_QP_PORT ? (attr->port_num - 1) : hr_qp->port;
+	active_mtu = iboe_get_mtu(hr_dev->iboe.netdevs[p]->mtu);
+
+	if ((hr_dev->caps.max_mtu >= IB_MTU_2048 &&
+	    attr->path_mtu > hr_dev->caps.max_mtu) ||
+	    attr->path_mtu < IB_MTU_256 || attr->path_mtu > active_mtu) {
+		ibdev_err(&hr_dev->ib_dev,
+			"attr path_mtu(%d)invalid while modify qp",
+			attr->path_mtu);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int hns_roce_check_qp_attr(struct ib_qp *ibqp, struct ib_qp_attr *attr,
+				  int attr_mask)
+{
+	struct hns_roce_dev *hr_dev = to_hr_dev(ibqp->device);
+	struct hns_roce_qp *hr_qp = to_hr_qp(ibqp);
+	int p;
+
+	if ((attr_mask & IB_QP_PORT) &&
+	    (attr->port_num == 0 || attr->port_num > hr_dev->caps.num_ports)) {
+		ibdev_err(&hr_dev->ib_dev,
+			"attr port_num invalid.attr->port_num=%d\n",
+			attr->port_num);
+		return -EINVAL;
+	}
+
+	if (attr_mask & IB_QP_PKEY_INDEX) {
+		p = attr_mask & IB_QP_PORT ? (attr->port_num - 1) : hr_qp->port;
+		if (attr->pkey_index >= hr_dev->caps.pkey_table_len[p]) {
+			ibdev_err(&hr_dev->ib_dev,
+				"attr pkey_index invalid.attr->pkey_index=%d\n",
+				attr->pkey_index);
+			return -EINVAL;
+		}
+	}
+
+	if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
+	    attr->max_rd_atomic > hr_dev->caps.max_qp_init_rdma) {
+		ibdev_err(&hr_dev->ib_dev,
+			"attr max_rd_atomic invalid.attr->max_rd_atomic=%d\n",
+			attr->max_rd_atomic);
+		return -EINVAL;
+	}
+
+	if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
+	    attr->max_dest_rd_atomic > hr_dev->caps.max_qp_dest_rdma) {
+		ibdev_err(&hr_dev->ib_dev,
+			"attr max_dest_rd_atomic invalid.attr->max_dest_rd_atomic=%d\n",
+			attr->max_dest_rd_atomic);
+		return -EINVAL;
+	}
+
+	if (attr_mask & IB_QP_PATH_MTU)
+		return check_mtu_validate(hr_dev, hr_qp, attr, attr_mask);
+
+	return 0;
+}
+
 int hns_roce_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		       int attr_mask, struct ib_udata *udata)
 {
 	struct hns_roce_dev *hr_dev = to_hr_dev(ibqp->device);
 	struct hns_roce_qp *hr_qp = to_hr_qp(ibqp);
 	enum ib_qp_state cur_state, new_state;
-	struct device *dev = hr_dev->dev;
 	int ret = -EINVAL;
-	int p;
-	enum ib_mtu active_mtu;
 
 	mutex_lock(&hr_qp->mutex);
 
 	cur_state = attr_mask & IB_QP_CUR_STATE ?
 		    attr->cur_qp_state : (enum ib_qp_state)hr_qp->state;
-	new_state = attr_mask & IB_QP_STATE ?
+	new_state = attr_mask & IB_QP_STATE ? attr->qp_state : cur_state;
-		    attr->qp_state : cur_state;
 
 	if (ibqp->uobject &&
 	    (attr_mask & IB_QP_STATE) && new_state == IB_QPS_ERR) {
@@ -1066,67 +1191,27 @@ int hns_roce_modify_qp(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 			if (hr_qp->rdb_en == 1)
 				hr_qp->rq.head = *(int *)(hr_qp->rdb.virt_addr);
 		} else {
-			dev_warn(dev, "flush cqe is not supported in userspace!\n");
+			ibdev_warn(&hr_dev->ib_dev,
+				  "flush cqe is not supported in userspace!\n");
 			goto out;
 		}
 	}
 
 	if (!ib_modify_qp_is_ok(cur_state, new_state, ibqp->qp_type,
 				attr_mask)) {
-		dev_err(dev, "ib_modify_qp_is_ok failed\n");
+		ibdev_err(&hr_dev->ib_dev, "ib_modify_qp_is_ok failed\n");
 		goto out;
 	}
 
-	if ((attr_mask & IB_QP_PORT) &&
+	ret = hns_roce_check_qp_attr(ibqp, attr, attr_mask);
-	    (attr->port_num == 0 || attr->port_num > hr_dev->caps.num_ports)) {
+	if (ret)
-		dev_err(dev, "attr port_num invalid.attr->port_num=%d\n",
-			attr->port_num);
 		goto out;
-	}
-
-	if (attr_mask & IB_QP_PKEY_INDEX) {
-		p = attr_mask & IB_QP_PORT ? (attr->port_num - 1) : hr_qp->port;
-		if (attr->pkey_index >= hr_dev->caps.pkey_table_len[p]) {
-			dev_err(dev, "attr pkey_index invalid.attr->pkey_index=%d\n",
-				attr->pkey_index);
-			goto out;
-		}
-	}
-
-	if (attr_mask & IB_QP_PATH_MTU) {
-		p = attr_mask & IB_QP_PORT ? (attr->port_num - 1) : hr_qp->port;
-		active_mtu = iboe_get_mtu(hr_dev->iboe.netdevs[p]->mtu);
-
-		if ((hr_dev->caps.max_mtu == IB_MTU_4096 &&
-		    attr->path_mtu > IB_MTU_4096) ||
-		    (hr_dev->caps.max_mtu == IB_MTU_2048 &&
-		    attr->path_mtu > IB_MTU_2048) ||
-		    attr->path_mtu < IB_MTU_256 ||
-		    attr->path_mtu > active_mtu) {
-			dev_err(dev, "attr path_mtu(%d)invalid while modify qp",
-				attr->path_mtu);
-			goto out;
-		}
-	}
-
-	if (attr_mask & IB_QP_MAX_QP_RD_ATOMIC &&
-	    attr->max_rd_atomic > hr_dev->caps.max_qp_init_rdma) {
-		dev_err(dev, "attr max_rd_atomic invalid.attr->max_rd_atomic=%d\n",
-			attr->max_rd_atomic);
-		goto out;
-	}
-
-	if (attr_mask & IB_QP_MAX_DEST_RD_ATOMIC &&
-	    attr->max_dest_rd_atomic > hr_dev->caps.max_qp_dest_rdma) {
-		dev_err(dev, "attr max_dest_rd_atomic invalid.attr->max_dest_rd_atomic=%d\n",
-			attr->max_dest_rd_atomic);
-		goto out;
-	}
 
 	if (cur_state == new_state && cur_state == IB_QPS_RESET) {
 		if (hr_dev->caps.min_wqes) {
 			ret = -EPERM;
-			dev_err(dev, "cur_state=%d new_state=%d\n", cur_state,
+			ibdev_err(&hr_dev->ib_dev,
+				"cur_state=%d new_state=%d\n", cur_state,
 				new_state);
 		} else {
 			ret = 0;

drivers/infiniband/hw/hns/hns_roce_srq.c

@@ -175,6 +175,76 @@ static void hns_roce_srq_free(struct hns_roce_dev *hr_dev,
 	hns_roce_bitmap_free(&srq_table->bitmap, srq->srqn, BITMAP_NO_RR);
 }
 
+static int create_user_srq(struct hns_roce_srq *srq, struct ib_udata *udata,
+			   int srq_buf_size)
+{
+	struct hns_roce_dev *hr_dev = to_hr_dev(srq->ibsrq.device);
+	struct hns_roce_ib_create_srq  ucmd;
+	u32 page_shift;
+	u32 npages;
+	int ret;
+
+	if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd)))
+		return -EFAULT;
+
+	srq->umem = ib_umem_get(udata, ucmd.buf_addr, srq_buf_size, 0, 0);
+	if (IS_ERR(srq->umem))
+		return PTR_ERR(srq->umem);
+
+	npages = (ib_umem_page_count(srq->umem) +
+		(1 << hr_dev->caps.srqwqe_buf_pg_sz) - 1) /
+		(1 << hr_dev->caps.srqwqe_buf_pg_sz);
+	page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
+	ret = hns_roce_mtt_init(hr_dev, npages, page_shift, &srq->mtt);
+	if (ret)
+		goto err_user_buf;
+
+	ret = hns_roce_ib_umem_write_mtt(hr_dev, &srq->mtt, srq->umem);
+	if (ret)
+		goto err_user_srq_mtt;
+
+	/* config index queue BA */
+	srq->idx_que.umem = ib_umem_get(udata, ucmd.que_addr,
+					srq->idx_que.buf_size, 0, 0);
+	if (IS_ERR(srq->idx_que.umem)) {
+		dev_err(hr_dev->dev, "ib_umem_get error for index queue\n");
+		ret = PTR_ERR(srq->idx_que.umem);
+		goto err_user_srq_mtt;
+	}
+
+	ret = hns_roce_mtt_init(hr_dev, ib_umem_page_count(srq->idx_que.umem),
+				PAGE_SHIFT, &srq->idx_que.mtt);
+
+	if (ret) {
+		dev_err(hr_dev->dev, "hns_roce_mtt_init error for idx que\n");
+		goto err_user_idx_mtt;
+	}
+
+	ret = hns_roce_ib_umem_write_mtt(hr_dev, &srq->idx_que.mtt,
+					 srq->idx_que.umem);
+	if (ret) {
+		dev_err(hr_dev->dev,
+			"hns_roce_ib_umem_write_mtt error for idx que\n");
+		goto err_user_idx_buf;
+	}
+
+	return 0;
+
+err_user_idx_buf:
+	hns_roce_mtt_cleanup(hr_dev, &srq->idx_que.mtt);
+
+err_user_idx_mtt:
+	ib_umem_release(srq->idx_que.umem);
+
+err_user_srq_mtt:
+	hns_roce_mtt_cleanup(hr_dev, &srq->mtt);
+
+err_user_buf:
+	ib_umem_release(srq->umem);
+
+	return ret;
+}
+
 static int hns_roce_create_idx_que(struct ib_pd *pd, struct hns_roce_srq *srq,
 				   u32 page_shift)
 {
@@ -196,6 +266,93 @@ static int hns_roce_create_idx_que(struct ib_pd *pd, struct hns_roce_srq *srq,
 	return 0;
 }
 
+static int create_kernel_srq(struct hns_roce_srq *srq, int srq_buf_size)
+{
+	struct hns_roce_dev *hr_dev = to_hr_dev(srq->ibsrq.device);
+	u32 page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
+	int ret;
+
+	if (hns_roce_buf_alloc(hr_dev, srq_buf_size, (1 << page_shift) * 2,
+			       &srq->buf, page_shift))
+		return -ENOMEM;
+
+	srq->head = 0;
+	srq->tail = srq->max - 1;
+
+	ret = hns_roce_mtt_init(hr_dev, srq->buf.npages, srq->buf.page_shift,
+				&srq->mtt);
+	if (ret)
+		goto err_kernel_buf;
+
+	ret = hns_roce_buf_write_mtt(hr_dev, &srq->mtt, &srq->buf);
+	if (ret)
+		goto err_kernel_srq_mtt;
+
+	page_shift = PAGE_SHIFT + hr_dev->caps.idx_buf_pg_sz;
+	ret = hns_roce_create_idx_que(srq->ibsrq.pd, srq, page_shift);
+	if (ret) {
+		dev_err(hr_dev->dev, "Create idx queue fail(%d)!\n", ret);
+		goto err_kernel_srq_mtt;
+	}
+
+	/* Init mtt table for idx_que */
+	ret = hns_roce_mtt_init(hr_dev, srq->idx_que.idx_buf.npages,
+				srq->idx_que.idx_buf.page_shift,
+				&srq->idx_que.mtt);
+	if (ret)
+		goto err_kernel_create_idx;
+
+	/* Write buffer address into the mtt table */
+	ret = hns_roce_buf_write_mtt(hr_dev, &srq->idx_que.mtt,
+				     &srq->idx_que.idx_buf);
+	if (ret)
+		goto err_kernel_idx_buf;
+
+	srq->wrid = kvmalloc_array(srq->max, sizeof(u64), GFP_KERNEL);
+	if (!srq->wrid) {
+		ret = -ENOMEM;
+		goto err_kernel_idx_buf;
+	}
+
+	return 0;
+
+err_kernel_idx_buf:
+	hns_roce_mtt_cleanup(hr_dev, &srq->idx_que.mtt);
+
+err_kernel_create_idx:
+	hns_roce_buf_free(hr_dev, srq->idx_que.buf_size,
+			  &srq->idx_que.idx_buf);
+	kfree(srq->idx_que.bitmap);
+
+err_kernel_srq_mtt:
+	hns_roce_mtt_cleanup(hr_dev, &srq->mtt);
+
+err_kernel_buf:
+	hns_roce_buf_free(hr_dev, srq_buf_size, &srq->buf);
+
+	return ret;
+}
+
+static void destroy_user_srq(struct hns_roce_dev *hr_dev,
+			     struct hns_roce_srq *srq)
+{
+	hns_roce_mtt_cleanup(hr_dev, &srq->idx_que.mtt);
+	ib_umem_release(srq->idx_que.umem);
+	hns_roce_mtt_cleanup(hr_dev, &srq->mtt);
+	ib_umem_release(srq->umem);
+}
+
+static void destroy_kernel_srq(struct hns_roce_dev *hr_dev,
+			       struct hns_roce_srq *srq, int srq_buf_size)
+{
+	kvfree(srq->wrid);
+	hns_roce_mtt_cleanup(hr_dev, &srq->idx_que.mtt);
+	hns_roce_buf_free(hr_dev, srq->idx_que.buf_size, &srq->idx_que.idx_buf);
+	kfree(srq->idx_que.bitmap);
+	hns_roce_mtt_cleanup(hr_dev, &srq->mtt);
+	hns_roce_buf_free(hr_dev, srq_buf_size, &srq->buf);
+}
+
 int hns_roce_create_srq(struct ib_srq *ib_srq,
 			struct ib_srq_init_attr *srq_init_attr,
 			struct ib_udata *udata)
@@ -205,9 +362,7 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
 	struct hns_roce_srq *srq = to_hr_srq(ib_srq);
 	int srq_desc_size;
 	int srq_buf_size;
-	u32 page_shift;
 	int ret = 0;
-	u32 npages;
 	u32 cqn;
 
 	/* Check the actual SRQ wqe and SRQ sge num */
@@ -233,115 +388,16 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
 	srq->idx_que.mtt.mtt_type = MTT_TYPE_IDX;
 
 	if (udata) {
-		struct hns_roce_ib_create_srq  ucmd;
+		ret = create_user_srq(srq, udata, srq_buf_size);
-
-		if (ib_copy_from_udata(&ucmd, udata, sizeof(ucmd)))
-			return -EFAULT;
-
-		srq->umem =
-			ib_umem_get(udata, ucmd.buf_addr, srq_buf_size, 0, 0);
-		if (IS_ERR(srq->umem))
-			return PTR_ERR(srq->umem);
-
-		if (hr_dev->caps.srqwqe_buf_pg_sz) {
-			npages = (ib_umem_page_count(srq->umem) +
-				  (1 << hr_dev->caps.srqwqe_buf_pg_sz) - 1) /
-				  (1 << hr_dev->caps.srqwqe_buf_pg_sz);
-			page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
-			ret = hns_roce_mtt_init(hr_dev, npages,
-						page_shift,
-						&srq->mtt);
-		} else
-			ret = hns_roce_mtt_init(hr_dev,
-						ib_umem_page_count(srq->umem),
-						PAGE_SHIFT, &srq->mtt);
-		if (ret)
-			goto err_buf;
-
-		ret = hns_roce_ib_umem_write_mtt(hr_dev, &srq->mtt, srq->umem);
-		if (ret)
-			goto err_srq_mtt;
-
-		/* config index queue BA */
-		srq->idx_que.umem = ib_umem_get(udata, ucmd.que_addr,
-						srq->idx_que.buf_size, 0, 0);
-		if (IS_ERR(srq->idx_que.umem)) {
-			dev_err(hr_dev->dev,
-				"ib_umem_get error for index queue\n");
-			ret = PTR_ERR(srq->idx_que.umem);
-			goto err_srq_mtt;
-		}
-
-		if (hr_dev->caps.idx_buf_pg_sz) {
-			npages = (ib_umem_page_count(srq->idx_que.umem) +
-				  (1 << hr_dev->caps.idx_buf_pg_sz) - 1) /
-				  (1 << hr_dev->caps.idx_buf_pg_sz);
-			page_shift = PAGE_SHIFT + hr_dev->caps.idx_buf_pg_sz;
-			ret = hns_roce_mtt_init(hr_dev, npages,
-						page_shift, &srq->idx_que.mtt);
-		} else {
-			ret = hns_roce_mtt_init(
-				hr_dev, ib_umem_page_count(srq->idx_que.umem),
-				PAGE_SHIFT, &srq->idx_que.mtt);
-		}
-
 		if (ret) {
-			dev_err(hr_dev->dev,
+			dev_err(hr_dev->dev, "Create user srq failed\n");
-				"hns_roce_mtt_init error for idx que\n");
+			goto err_srq;
-			goto err_idx_mtt;
-		}
-
-		ret = hns_roce_ib_umem_write_mtt(hr_dev, &srq->idx_que.mtt,
-						 srq->idx_que.umem);
-		if (ret) {
-			dev_err(hr_dev->dev,
-			      "hns_roce_ib_umem_write_mtt error for idx que\n");
-			goto err_idx_buf;
 		}
 	} else {
-		page_shift = PAGE_SHIFT + hr_dev->caps.srqwqe_buf_pg_sz;
+		ret = create_kernel_srq(srq, srq_buf_size);
-		if (hns_roce_buf_alloc(hr_dev, srq_buf_size,
-				       (1 << page_shift) * 2, &srq->buf,
-				       page_shift))
-			return -ENOMEM;
-
-		srq->head = 0;
-		srq->tail = srq->max - 1;
-
-		ret = hns_roce_mtt_init(hr_dev, srq->buf.npages,
-					srq->buf.page_shift, &srq->mtt);
-		if (ret)
-			goto err_buf;
-
-		ret = hns_roce_buf_write_mtt(hr_dev, &srq->mtt, &srq->buf);
-		if (ret)
-			goto err_srq_mtt;
-
-		page_shift = PAGE_SHIFT + hr_dev->caps.idx_buf_pg_sz;
-		ret = hns_roce_create_idx_que(ib_srq->pd, srq, page_shift);
 		if (ret) {
-			dev_err(hr_dev->dev, "Create idx queue fail(%d)!\n",
+			dev_err(hr_dev->dev, "Create kernel srq failed\n");
-				ret);
+			goto err_srq;
-			goto err_srq_mtt;
-		}
-
-		/* Init mtt table for idx_que */
-		ret = hns_roce_mtt_init(hr_dev, srq->idx_que.idx_buf.npages,
-					srq->idx_que.idx_buf.page_shift,
-					&srq->idx_que.mtt);
-		if (ret)
-			goto err_create_idx;
-
-		/* Write buffer address into the mtt table */
-		ret = hns_roce_buf_write_mtt(hr_dev, &srq->idx_que.mtt,
-					     &srq->idx_que.idx_buf);
-		if (ret)
-			goto err_idx_buf;
-
-		srq->wrid = kvmalloc_array(srq->max, sizeof(u64), GFP_KERNEL);
-		if (!srq->wrid) {
-			ret = -ENOMEM;
-			goto err_idx_buf;
 		}
 	}
 
@@ -356,7 +412,6 @@ int hns_roce_create_srq(struct ib_srq *ib_srq,
 		goto err_wrid;
 
 	srq->event = hns_roce_ib_srq_event;
-	srq->ibsrq.ext.xrc.srq_num = srq->srqn;
 	resp.srqn = srq->srqn;
 
 	if (udata) {
@@ -373,27 +428,12 @@ err_srqc_alloc:
 	hns_roce_srq_free(hr_dev, srq);
 
 err_wrid:
-	kvfree(srq->wrid);
+	if (udata)
-
+		destroy_user_srq(hr_dev, srq);
-err_idx_buf:
+	else
-	hns_roce_mtt_cleanup(hr_dev, &srq->idx_que.mtt);
+		destroy_kernel_srq(hr_dev, srq, srq_buf_size);
-
-err_idx_mtt:
-	ib_umem_release(srq->idx_que.umem);
-
-err_create_idx:
-	hns_roce_buf_free(hr_dev, srq->idx_que.buf_size,
-			  &srq->idx_que.idx_buf);
-	bitmap_free(srq->idx_que.bitmap);
-
-err_srq_mtt:
-	hns_roce_mtt_cleanup(hr_dev, &srq->mtt);
-
-err_buf:
-	ib_umem_release(srq->umem);
-	if (!udata)
-		hns_roce_buf_free(hr_dev, srq_buf_size, &srq->buf);
 
+err_srq:
 	return ret;
 }
 

drivers/infiniband/hw/i40iw/i40iw_verbs.c

@@ -97,18 +97,7 @@ static int i40iw_query_port(struct ib_device *ibdev,
 			    u8 port,
 			    struct ib_port_attr *props)
 {
-	struct i40iw_device *iwdev = to_iwdev(ibdev);
-	struct net_device *netdev = iwdev->netdev;
-
-	/* props being zeroed by the caller, avoid zeroing it here */
-	props->max_mtu = IB_MTU_4096;
-	props->active_mtu = ib_mtu_int_to_enum(netdev->mtu);
-
 	props->lid = 1;
-	if (netif_carrier_ok(iwdev->netdev))
-		props->state = IB_PORT_ACTIVE;
-	else
-		props->state = IB_PORT_DOWN;
 	props->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_REINIT_SUP |
 		IB_PORT_VENDOR_CLASS_SUP | IB_PORT_BOOT_MGMT_SUP;
 	props->gid_tbl_len = 1;

drivers/infiniband/hw/mlx4/main.c

@@ -734,7 +734,8 @@ out:
 
 static u8 state_to_phys_state(enum ib_port_state state)
 {
-	return state == IB_PORT_ACTIVE ? 5 : 3;
+	return state == IB_PORT_ACTIVE ?
+		IB_PORT_PHYS_STATE_LINK_UP : IB_PORT_PHYS_STATE_DISABLED;
 }
 
 static int eth_link_query_port(struct ib_device *ibdev, u8 port,

drivers/infiniband/hw/mlx4/mr.c

@@ -377,6 +377,7 @@ static struct ib_umem *mlx4_get_umem_mr(struct ib_udata *udata, u64 start,
 	 * again
 	 */
 	if (!ib_access_writable(access_flags)) {
+		unsigned long untagged_start = untagged_addr(start);
 		struct vm_area_struct *vma;
 
 		down_read(&current->mm->mmap_sem);
@@ -385,9 +386,9 @@ static struct ib_umem *mlx4_get_umem_mr(struct ib_udata *udata, u64 start,
 		 * cover the memory, but for now it requires a single vma to
 		 * entirely cover the MR to support RO mappings.
 		 */
-		vma = find_vma(current->mm, start);
+		vma = find_vma(current->mm, untagged_start);
-		if (vma && vma->vm_end >= start + length &&
+		if (vma && vma->vm_end >= untagged_start + length &&
-		    vma->vm_start <= start) {
+		    vma->vm_start <= untagged_start) {
 			if (vma->vm_flags & VM_WRITE)
 				access_flags |= IB_ACCESS_LOCAL_WRITE;
 		} else {

drivers/infiniband/hw/mlx4/qp.c

@@ -325,7 +325,7 @@ static int send_wqe_overhead(enum mlx4_ib_qp_type type, u32 flags)
 }
 
 static int set_rq_size(struct mlx4_ib_dev *dev, struct ib_qp_cap *cap,
-		       bool is_user, int has_rq, struct mlx4_ib_qp *qp,
+		       bool is_user, bool has_rq, struct mlx4_ib_qp *qp,
 		       u32 inl_recv_sz)
 {
 	/* Sanity check RQ size before proceeding */
@@ -506,10 +506,10 @@ static void free_proxy_bufs(struct ib_device *dev, struct mlx4_ib_qp *qp)
 	kfree(qp->sqp_proxy_rcv);
 }
 
-static int qp_has_rq(struct ib_qp_init_attr *attr)
+static bool qp_has_rq(struct ib_qp_init_attr *attr)
 {
 	if (attr->qp_type == IB_QPT_XRC_INI || attr->qp_type == IB_QPT_XRC_TGT)
-		return 0;
+		return false;
 
 	return !attr->srq;
 }
@@ -855,12 +855,143 @@ static void mlx4_ib_release_wqn(struct mlx4_ib_ucontext *context,
 	mutex_unlock(&context->wqn_ranges_mutex);
 }
 
-static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
+static int create_rq(struct ib_pd *pd, struct ib_qp_init_attr *init_attr,
-			    enum mlx4_ib_source_type src,
+		     struct ib_udata *udata, struct mlx4_ib_qp *qp)
-			    struct ib_qp_init_attr *init_attr,
+{
+	struct mlx4_ib_dev *dev = to_mdev(pd->device);
+	int qpn;
+	int err;
+	struct mlx4_ib_ucontext *context = rdma_udata_to_drv_context(
+		udata, struct mlx4_ib_ucontext, ibucontext);
+	struct mlx4_ib_cq *mcq;
+	unsigned long flags;
+	int range_size;
+	struct mlx4_ib_create_wq wq;
+	size_t copy_len;
+	int shift;
+	int n;
+
+	qp->mlx4_ib_qp_type = MLX4_IB_QPT_RAW_PACKET;
+
+	mutex_init(&qp->mutex);
+	spin_lock_init(&qp->sq.lock);
+	spin_lock_init(&qp->rq.lock);
+	INIT_LIST_HEAD(&qp->gid_list);
+	INIT_LIST_HEAD(&qp->steering_rules);
+
+	qp->state = IB_QPS_RESET;
+
+	copy_len = min(sizeof(struct mlx4_ib_create_wq), udata->inlen);
+
+	if (ib_copy_from_udata(&wq, udata, copy_len)) {
+		err = -EFAULT;
+		goto err;
+	}
+
+	if (wq.comp_mask || wq.reserved[0] || wq.reserved[1] ||
+	    wq.reserved[2]) {
+		pr_debug("user command isn't supported\n");
+		err = -EOPNOTSUPP;
+		goto err;
+	}
+
+	if (wq.log_range_size > ilog2(dev->dev->caps.max_rss_tbl_sz)) {
+		pr_debug("WQN range size must be equal or smaller than %d\n",
+			 dev->dev->caps.max_rss_tbl_sz);
+		err = -EOPNOTSUPP;
+		goto err;
+	}
+	range_size = 1 << wq.log_range_size;
+
+	if (init_attr->create_flags & IB_QP_CREATE_SCATTER_FCS)
+		qp->flags |= MLX4_IB_QP_SCATTER_FCS;
+
+	err = set_rq_size(dev, &init_attr->cap, true, true, qp, qp->inl_recv_sz);
+	if (err)
+		goto err;
+
+	qp->sq_no_prefetch = 1;
+	qp->sq.wqe_cnt = 1;
+	qp->sq.wqe_shift = MLX4_IB_MIN_SQ_STRIDE;
+	qp->buf_size = (qp->rq.wqe_cnt << qp->rq.wqe_shift) +
+		       (qp->sq.wqe_cnt << qp->sq.wqe_shift);
+
+	qp->umem = ib_umem_get(udata, wq.buf_addr, qp->buf_size, 0, 0);
+	if (IS_ERR(qp->umem)) {
+		err = PTR_ERR(qp->umem);
+		goto err;
+	}
+
+	n = ib_umem_page_count(qp->umem);
+	shift = mlx4_ib_umem_calc_optimal_mtt_size(qp->umem, 0, &n);
+	err = mlx4_mtt_init(dev->dev, n, shift, &qp->mtt);
+
+	if (err)
+		goto err_buf;
+
+	err = mlx4_ib_umem_write_mtt(dev, &qp->mtt, qp->umem);
+	if (err)
+		goto err_mtt;
+
+	err = mlx4_ib_db_map_user(udata, wq.db_addr, &qp->db);
+	if (err)
+		goto err_mtt;
+	qp->mqp.usage = MLX4_RES_USAGE_USER_VERBS;
+
+	err = mlx4_ib_alloc_wqn(context, qp, range_size, &qpn);
+	if (err)
+		goto err_wrid;
+
+	err = mlx4_qp_alloc(dev->dev, qpn, &qp->mqp);
+	if (err)
+		goto err_qpn;
+
+	/*
+	 * Hardware wants QPN written in big-endian order (after
+	 * shifting) for send doorbell.  Precompute this value to save
+	 * a little bit when posting sends.
+	 */
+	qp->doorbell_qpn = swab32(qp->mqp.qpn << 8);
+
+	qp->mqp.event = mlx4_ib_wq_event;
+
+	spin_lock_irqsave(&dev->reset_flow_resource_lock, flags);
+	mlx4_ib_lock_cqs(to_mcq(init_attr->send_cq),
+			 to_mcq(init_attr->recv_cq));
+	/* Maintain device to QPs access, needed for further handling
+	 * via reset flow
+	 */
+	list_add_tail(&qp->qps_list, &dev->qp_list);
+	/* Maintain CQ to QPs access, needed for further handling
+	 * via reset flow
+	 */
+	mcq = to_mcq(init_attr->send_cq);
+	list_add_tail(&qp->cq_send_list, &mcq->send_qp_list);
+	mcq = to_mcq(init_attr->recv_cq);
+	list_add_tail(&qp->cq_recv_list, &mcq->recv_qp_list);
+	mlx4_ib_unlock_cqs(to_mcq(init_attr->send_cq),
+			   to_mcq(init_attr->recv_cq));
+	spin_unlock_irqrestore(&dev->reset_flow_resource_lock, flags);
+	return 0;
+
+err_qpn:
+	mlx4_ib_release_wqn(context, qp, 0);
+err_wrid:
+	mlx4_ib_db_unmap_user(context, &qp->db);
+
+err_mtt:
+	mlx4_mtt_cleanup(dev->dev, &qp->mtt);
+err_buf:
+	ib_umem_release(qp->umem);
+err:
+	return err;
+}
+
+static int create_qp_common(struct ib_pd *pd, struct ib_qp_init_attr *init_attr,
 			    struct ib_udata *udata, int sqpn,
 			    struct mlx4_ib_qp **caller_qp)
 {
+	struct mlx4_ib_dev *dev = to_mdev(pd->device);
 	int qpn;
 	int err;
 	struct mlx4_ib_sqp *sqp = NULL;
@@ -870,7 +1001,6 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 	enum mlx4_ib_qp_type qp_type = (enum mlx4_ib_qp_type) init_attr->qp_type;
 	struct mlx4_ib_cq *mcq;
 	unsigned long flags;
-	int range_size = 0;
 
 	/* When tunneling special qps, we use a plain UD qp */
 	if (sqpn) {
@@ -921,15 +1051,13 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 			if (!sqp)
 				return -ENOMEM;
 			qp = &sqp->qp;
-			qp->pri.vid = 0xFFFF;
-			qp->alt.vid = 0xFFFF;
 		} else {
 			qp = kzalloc(sizeof(struct mlx4_ib_qp), GFP_KERNEL);
 			if (!qp)
 				return -ENOMEM;
-			qp->pri.vid = 0xFFFF;
-			qp->alt.vid = 0xFFFF;
 		}
+		qp->pri.vid = 0xFFFF;
+		qp->alt.vid = 0xFFFF;
 	} else
 		qp = *caller_qp;
 
@@ -941,48 +1069,24 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 	INIT_LIST_HEAD(&qp->gid_list);
 	INIT_LIST_HEAD(&qp->steering_rules);
 
-	qp->state	 = IB_QPS_RESET;
+	qp->state = IB_QPS_RESET;
 	if (init_attr->sq_sig_type == IB_SIGNAL_ALL_WR)
 		qp->sq_signal_bits = cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE);
 
-
 	if (udata) {
-		union {
+		struct mlx4_ib_create_qp ucmd;
-			struct mlx4_ib_create_qp qp;
-			struct mlx4_ib_create_wq wq;
-		} ucmd;
 		size_t copy_len;
 		int shift;
 		int n;
 
-		copy_len = (src == MLX4_IB_QP_SRC) ?
+		copy_len = sizeof(struct mlx4_ib_create_qp);
-			   sizeof(struct mlx4_ib_create_qp) :
-			   min(sizeof(struct mlx4_ib_create_wq), udata->inlen);
 
 		if (ib_copy_from_udata(&ucmd, udata, copy_len)) {
 			err = -EFAULT;
 			goto err;
 		}
 
-		if (src == MLX4_IB_RWQ_SRC) {
+		qp->inl_recv_sz = ucmd.inl_recv_sz;
-			if (ucmd.wq.comp_mask || ucmd.wq.reserved[0] ||
-			    ucmd.wq.reserved[1] || ucmd.wq.reserved[2]) {
-				pr_debug("user command isn't supported\n");
-				err = -EOPNOTSUPP;
-				goto err;
-			}
-
-			if (ucmd.wq.log_range_size >
-			    ilog2(dev->dev->caps.max_rss_tbl_sz)) {
-				pr_debug("WQN range size must be equal or smaller than %d\n",
-					 dev->dev->caps.max_rss_tbl_sz);
-				err = -EOPNOTSUPP;
-				goto err;
-			}
-			range_size = 1 << ucmd.wq.log_range_size;
-		} else {
-			qp->inl_recv_sz = ucmd.qp.inl_recv_sz;
-		}
 
 		if (init_attr->create_flags & IB_QP_CREATE_SCATTER_FCS) {
 			if (!(dev->dev->caps.flags &
@@ -1000,30 +1104,14 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 		if (err)
 			goto err;
 
-		if (src == MLX4_IB_QP_SRC) {
+		qp->sq_no_prefetch = ucmd.sq_no_prefetch;
-			qp->sq_no_prefetch = ucmd.qp.sq_no_prefetch;
 
-			err = set_user_sq_size(dev, qp,
+		err = set_user_sq_size(dev, qp, &ucmd);
-					       (struct mlx4_ib_create_qp *)
+		if (err)
-					       &ucmd);
+			goto err;
-			if (err)
-				goto err;
-		} else {
-			qp->sq_no_prefetch = 1;
-			qp->sq.wqe_cnt = 1;
-			qp->sq.wqe_shift = MLX4_IB_MIN_SQ_STRIDE;
-			/* Allocated buffer expects to have at least that SQ
-			 * size.
-			 */
-			qp->buf_size = (qp->rq.wqe_cnt << qp->rq.wqe_shift) +
-				(qp->sq.wqe_cnt << qp->sq.wqe_shift);
-		}
 
 		qp->umem =
-			ib_umem_get(udata,
+			ib_umem_get(udata, ucmd.buf_addr, qp->buf_size, 0, 0);
-				    (src == MLX4_IB_QP_SRC) ? ucmd.qp.buf_addr :
-							      ucmd.wq.buf_addr,
-				    qp->buf_size, 0, 0);
 		if (IS_ERR(qp->umem)) {
 			err = PTR_ERR(qp->umem);
 			goto err;
@@ -1041,11 +1129,7 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 			goto err_mtt;
 
 		if (qp_has_rq(init_attr)) {
-			err = mlx4_ib_db_map_user(udata,
+			err = mlx4_ib_db_map_user(udata, ucmd.db_addr, &qp->db);
-						  (src == MLX4_IB_QP_SRC) ?
-							  ucmd.qp.db_addr :
-							  ucmd.wq.db_addr,
-						  &qp->db);
 			if (err)
 				goto err_mtt;
 		}
@@ -1115,10 +1199,6 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 				goto err_wrid;
 			}
 		}
-	} else if (src == MLX4_IB_RWQ_SRC) {
-		err = mlx4_ib_alloc_wqn(context, qp, range_size, &qpn);
-		if (err)
-			goto err_wrid;
 	} else {
 		/* Raw packet QPNs may not have bits 6,7 set in their qp_num;
 		 * otherwise, the WQE BlueFlame setup flow wrongly causes
@@ -1157,8 +1237,7 @@ static int create_qp_common(struct mlx4_ib_dev *dev, struct ib_pd *pd,
 	 */
 	qp->doorbell_qpn = swab32(qp->mqp.qpn << 8);
 
-	qp->mqp.event = (src == MLX4_IB_QP_SRC) ? mlx4_ib_qp_event :
+	qp->mqp.event = mlx4_ib_qp_event;
-						  mlx4_ib_wq_event;
 
 	if (!*caller_qp)
 		*caller_qp = qp;
@@ -1186,8 +1265,6 @@ err_qpn:
 	if (!sqpn) {
 		if (qp->flags & MLX4_IB_QP_NETIF)
 			mlx4_ib_steer_qp_free(dev, qpn, 1);
-		else if (src == MLX4_IB_RWQ_SRC)
-			mlx4_ib_release_wqn(context, qp, 0);
 		else
 			mlx4_qp_release_range(dev->dev, qpn, 1);
 	}
@@ -1518,8 +1595,7 @@ static struct ib_qp *_mlx4_ib_create_qp(struct ib_pd *pd,
 		/* fall through */
 	case IB_QPT_UD:
 	{
-		err = create_qp_common(to_mdev(pd->device), pd,	MLX4_IB_QP_SRC,
+		err = create_qp_common(pd, init_attr, udata, 0, &qp);
-				       init_attr, udata, 0, &qp);
 		if (err) {
 			kfree(qp);
 			return ERR_PTR(err);
@@ -1549,8 +1625,7 @@ static struct ib_qp *_mlx4_ib_create_qp(struct ib_pd *pd,
 			sqpn = get_sqp_num(to_mdev(pd->device), init_attr);
 		}
 
-		err = create_qp_common(to_mdev(pd->device), pd, MLX4_IB_QP_SRC,
+		err = create_qp_common(pd, init_attr, udata, sqpn, &qp);
-				       init_attr, udata, sqpn, &qp);
 		if (err)
 			return ERR_PTR(err);
 
@@ -4047,8 +4122,8 @@ struct ib_wq *mlx4_ib_create_wq(struct ib_pd *pd,
 				struct ib_wq_init_attr *init_attr,
 				struct ib_udata *udata)
 {
-	struct mlx4_ib_dev *dev;
+	struct mlx4_dev *dev = to_mdev(pd->device)->dev;
-	struct ib_qp_init_attr ib_qp_init_attr;
+	struct ib_qp_init_attr ib_qp_init_attr = {};
 	struct mlx4_ib_qp *qp;
 	struct mlx4_ib_create_wq ucmd;
 	int err, required_cmd_sz;
@@ -4073,14 +4148,13 @@ struct ib_wq *mlx4_ib_create_wq(struct ib_pd *pd,
 	if (udata->outlen)
 		return ERR_PTR(-EOPNOTSUPP);
 
-	dev = to_mdev(pd->device);
-
 	if (init_attr->wq_type != IB_WQT_RQ) {
 		pr_debug("unsupported wq type %d\n", init_attr->wq_type);
 		return ERR_PTR(-EOPNOTSUPP);
 	}
 
-	if (init_attr->create_flags & ~IB_WQ_FLAGS_SCATTER_FCS) {
+	if (init_attr->create_flags & ~IB_WQ_FLAGS_SCATTER_FCS ||
+	    !(dev->caps.flags & MLX4_DEV_CAP_FLAG_FCS_KEEP)) {
 		pr_debug("unsupported create_flags %u\n",
 			 init_attr->create_flags);
 		return ERR_PTR(-EOPNOTSUPP);
@@ -4093,7 +4167,6 @@ struct ib_wq *mlx4_ib_create_wq(struct ib_pd *pd,
 	qp->pri.vid = 0xFFFF;
 	qp->alt.vid = 0xFFFF;
 
-	memset(&ib_qp_init_attr, 0, sizeof(ib_qp_init_attr));
 	ib_qp_init_attr.qp_context = init_attr->wq_context;
 	ib_qp_init_attr.qp_type = IB_QPT_RAW_PACKET;
 	ib_qp_init_attr.cap.max_recv_wr = init_attr->max_wr;
@@ -4104,8 +4177,7 @@ struct ib_wq *mlx4_ib_create_wq(struct ib_pd *pd,
 	if (init_attr->create_flags & IB_WQ_FLAGS_SCATTER_FCS)
 		ib_qp_init_attr.create_flags |= IB_QP_CREATE_SCATTER_FCS;
 
-	err = create_qp_common(dev, pd, MLX4_IB_RWQ_SRC, &ib_qp_init_attr,
+	err = create_rq(pd, &ib_qp_init_attr, udata, qp);
-			       udata, 0, &qp);
 	if (err) {
 		kfree(qp);
 		return ERR_PTR(err);

drivers/infiniband/hw/mlx5/devx.c

@@ -233,6 +233,8 @@ static bool is_legacy_obj_event_num(u16 event_num)
 	case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR:
 	case MLX5_EVENT_TYPE_DCT_DRAINED:
 	case MLX5_EVENT_TYPE_COMP:
+	case MLX5_EVENT_TYPE_DCT_KEY_VIOLATION:
+	case MLX5_EVENT_TYPE_XRQ_ERROR:
 		return true;
 	default:
 		return false;
@@ -315,8 +317,10 @@ static u16 get_event_obj_type(unsigned long event_type, struct mlx5_eqe *eqe)
 	case MLX5_EVENT_TYPE_SRQ_CATAS_ERROR:
 		return eqe->data.qp_srq.type;
 	case MLX5_EVENT_TYPE_CQ_ERROR:
+	case MLX5_EVENT_TYPE_XRQ_ERROR:
 		return 0;
 	case MLX5_EVENT_TYPE_DCT_DRAINED:
+	case MLX5_EVENT_TYPE_DCT_KEY_VIOLATION:
 		return MLX5_EVENT_QUEUE_TYPE_DCT;
 	default:
 		return MLX5_GET(affiliated_event_header, &eqe->data, obj_type);
@@ -542,6 +546,8 @@ static u64 devx_get_obj_id(const void *in)
 		break;
 	case MLX5_CMD_OP_ARM_XRQ:
 	case MLX5_CMD_OP_SET_XRQ_DC_PARAMS_ENTRY:
+	case MLX5_CMD_OP_RELEASE_XRQ_ERROR:
+	case MLX5_CMD_OP_MODIFY_XRQ:
 		obj_id = get_enc_obj_id(MLX5_CMD_OP_CREATE_XRQ,
 					MLX5_GET(arm_xrq_in, in, xrqn));
 		break;
@@ -776,6 +782,14 @@ static bool devx_is_obj_create_cmd(const void *in, u16 *opcode)
 			return true;
 		return false;
 	}
+	case MLX5_CMD_OP_CREATE_PSV:
+	{
+		u8 num_psv = MLX5_GET(create_psv_in, in, num_psv);
+
+		if (num_psv == 1)
+			return true;
+		return false;
+	}
 	default:
 		return false;
 	}
@@ -810,6 +824,8 @@ static bool devx_is_obj_modify_cmd(const void *in)
 	case MLX5_CMD_OP_ARM_DCT_FOR_KEY_VIOLATION:
 	case MLX5_CMD_OP_ARM_XRQ:
 	case MLX5_CMD_OP_SET_XRQ_DC_PARAMS_ENTRY:
+	case MLX5_CMD_OP_RELEASE_XRQ_ERROR:
+	case MLX5_CMD_OP_MODIFY_XRQ:
 		return true;
 	case MLX5_CMD_OP_SET_FLOW_TABLE_ENTRY:
 	{
@@ -1216,6 +1232,12 @@ static void devx_obj_build_destroy_cmd(void *in, void *out, void *din,
 	case MLX5_CMD_OP_ALLOC_XRCD:
 		MLX5_SET(general_obj_in_cmd_hdr, din, opcode, MLX5_CMD_OP_DEALLOC_XRCD);
 		break;
+	case MLX5_CMD_OP_CREATE_PSV:
+		MLX5_SET(general_obj_in_cmd_hdr, din, opcode,
+			 MLX5_CMD_OP_DESTROY_PSV);
+		MLX5_SET(destroy_psv_in, din, psvn,
+			 MLX5_GET(create_psv_out, out, psv0_index));
+		break;
 	default:
 		/* The entry must match to one of the devx_is_obj_create_cmd */
 		WARN_ON(true);
@@ -2286,7 +2308,11 @@ static u32 devx_get_obj_id_from_event(unsigned long event_type, void *data)
 	case MLX5_EVENT_TYPE_WQ_ACCESS_ERROR:
 		obj_id = be32_to_cpu(eqe->data.qp_srq.qp_srq_n) & 0xffffff;
 		break;
+	case MLX5_EVENT_TYPE_XRQ_ERROR:
+		obj_id = be32_to_cpu(eqe->data.xrq_err.type_xrqn) & 0xffffff;
+		break;
 	case MLX5_EVENT_TYPE_DCT_DRAINED:
+	case MLX5_EVENT_TYPE_DCT_KEY_VIOLATION:
 		obj_id = be32_to_cpu(eqe->data.dct.dctn) & 0xffffff;
 		break;
 	case MLX5_EVENT_TYPE_CQ_ERROR:

drivers/infiniband/hw/mlx5/flow.c

@@ -32,6 +32,9 @@ mlx5_ib_ft_type_to_namespace(enum mlx5_ib_uapi_flow_table_type table_type,
 	case MLX5_IB_UAPI_FLOW_TABLE_TYPE_FDB:
 		*namespace = MLX5_FLOW_NAMESPACE_FDB;
 		break;
+	case MLX5_IB_UAPI_FLOW_TABLE_TYPE_RDMA_RX:
+		*namespace = MLX5_FLOW_NAMESPACE_RDMA_RX;
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -101,6 +104,11 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
 	if (fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_FDB && !dest_devx)
 		return -EINVAL;
 
+	/* Allow only DEVX object or QP as dest when inserting to RDMA_RX */
+	if ((fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_RDMA_RX) &&
+	    ((!dest_devx && !dest_qp) || (dest_devx && dest_qp)))
+		return -EINVAL;
+
 	if (dest_devx) {
 		devx_obj = uverbs_attr_get_obj(
 			attrs, MLX5_IB_ATTR_CREATE_FLOW_DEST_DEVX);
@@ -112,8 +120,9 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
 		 */
 		if (!mlx5_ib_devx_is_flow_dest(devx_obj, &dest_id, &dest_type))
 			return -EINVAL;
-		/* Allow only flow table as dest when inserting to FDB */
+		/* Allow only flow table as dest when inserting to FDB or RDMA_RX */
-		if (fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_FDB &&
+		if ((fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_FDB ||
+		     fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_RDMA_RX) &&
 		    dest_type != MLX5_FLOW_DESTINATION_TYPE_FLOW_TABLE)
 			return -EINVAL;
 	} else if (dest_qp) {

drivers/infiniband/hw/mlx5/main.c

@@ -535,7 +535,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
 	props->max_msg_sz       = 1 << MLX5_CAP_GEN(dev->mdev, log_max_msg);
 	props->pkey_tbl_len     = 1;
 	props->state            = IB_PORT_DOWN;
-	props->phys_state       = 3;
+	props->phys_state       = IB_PORT_PHYS_STATE_DISABLED;
 
 	mlx5_query_nic_vport_qkey_viol_cntr(mdev, &qkey_viol_cntr);
 	props->qkey_viol_cntr = qkey_viol_cntr;
@@ -561,7 +561,7 @@ static int mlx5_query_port_roce(struct ib_device *device, u8 port_num,
 
 	if (netif_running(ndev) && netif_carrier_ok(ndev)) {
 		props->state      = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	}
 
 	ndev_ib_mtu = iboe_get_mtu(ndev->mtu);
@@ -3971,6 +3971,11 @@ _get_flow_table(struct mlx5_ib_dev *dev,
 		    esw_encap)
 			flags |= MLX5_FLOW_TABLE_TUNNEL_EN_REFORMAT;
 		priority = FDB_BYPASS_PATH;
+	} else if (fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_RDMA_RX) {
+		max_table_size =
+			BIT(MLX5_CAP_FLOWTABLE_RDMA_RX(dev->mdev,
+						       log_max_ft_size));
+		priority = fs_matcher->priority;
 	}
 
 	max_table_size = min_t(int, max_table_size, MLX5_FS_MAX_ENTRIES);
@@ -3985,6 +3990,8 @@ _get_flow_table(struct mlx5_ib_dev *dev,
 		prio = &dev->flow_db->egress_prios[priority];
 	else if (fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_FDB)
 		prio = &dev->flow_db->fdb;
+	else if (fs_matcher->ns_type == MLX5_FLOW_NAMESPACE_RDMA_RX)
+		prio = &dev->flow_db->rdma_rx[priority];
 
 	if (!prio)
 		return ERR_PTR(-EINVAL);
@@ -5326,11 +5333,21 @@ static const struct mlx5_ib_counter ext_ppcnt_cnts[] = {
 	INIT_EXT_PPCNT_COUNTER(rx_icrc_encapsulated),
 };
 
+static bool is_mdev_switchdev_mode(const struct mlx5_core_dev *mdev)
+{
+	return MLX5_ESWITCH_MANAGER(mdev) &&
+	       mlx5_ib_eswitch_mode(mdev->priv.eswitch) ==
+		       MLX5_ESWITCH_OFFLOADS;
+}
+
 static void mlx5_ib_dealloc_counters(struct mlx5_ib_dev *dev)
 {
+	int num_cnt_ports;
 	int i;
 
-	for (i = 0; i < dev->num_ports; i++) {
+	num_cnt_ports = is_mdev_switchdev_mode(dev->mdev) ? 1 : dev->num_ports;
+
+	for (i = 0; i < num_cnt_ports; i++) {
 		if (dev->port[i].cnts.set_id_valid)
 			mlx5_core_dealloc_q_counter(dev->mdev,
 						    dev->port[i].cnts.set_id);
@@ -5432,13 +5449,15 @@ static void mlx5_ib_fill_counters(struct mlx5_ib_dev *dev,
 
 static int mlx5_ib_alloc_counters(struct mlx5_ib_dev *dev)
 {
+	int num_cnt_ports;
 	int err = 0;
 	int i;
 	bool is_shared;
 
 	is_shared = MLX5_CAP_GEN(dev->mdev, log_max_uctx) != 0;
+	num_cnt_ports = is_mdev_switchdev_mode(dev->mdev) ? 1 : dev->num_ports;
 
-	for (i = 0; i < dev->num_ports; i++) {
+	for (i = 0; i < num_cnt_ports; i++) {
 		err = __mlx5_ib_alloc_counters(dev, &dev->port[i].cnts);
 		if (err)
 			goto err_alloc;
@@ -5458,7 +5477,6 @@ static int mlx5_ib_alloc_counters(struct mlx5_ib_dev *dev)
 		}
 		dev->port[i].cnts.set_id_valid = true;
 	}
-
 	return 0;
 
 err_alloc:
@@ -5466,25 +5484,50 @@ err_alloc:
 	return err;
 }
 
+static const struct mlx5_ib_counters *get_counters(struct mlx5_ib_dev *dev,
+						   u8 port_num)
+{
+	return is_mdev_switchdev_mode(dev->mdev) ? &dev->port[0].cnts :
+						   &dev->port[port_num].cnts;
+}
+
+/**
+ * mlx5_ib_get_counters_id - Returns counters id to use for device+port
+ * @dev:	Pointer to mlx5 IB device
+ * @port_num:	Zero based port number
+ *
+ * mlx5_ib_get_counters_id() Returns counters set id to use for given
+ * device port combination in switchdev and non switchdev mode of the
+ * parent device.
+ */
+u16 mlx5_ib_get_counters_id(struct mlx5_ib_dev *dev, u8 port_num)
+{
+	const struct mlx5_ib_counters *cnts = get_counters(dev, port_num);
+
+	return cnts->set_id;
+}
+
 static struct rdma_hw_stats *mlx5_ib_alloc_hw_stats(struct ib_device *ibdev,
 						    u8 port_num)
 {
 	struct mlx5_ib_dev *dev = to_mdev(ibdev);
-	struct mlx5_ib_port *port = &dev->port[port_num - 1];
+	const struct mlx5_ib_counters *cnts;
+	bool is_switchdev = is_mdev_switchdev_mode(dev->mdev);
 
-	/* We support only per port stats */
+	if ((is_switchdev && port_num) || (!is_switchdev && !port_num))
-	if (port_num == 0)
 		return NULL;
 
-	return rdma_alloc_hw_stats_struct(port->cnts.names,
+	cnts = get_counters(dev, port_num - 1);
-					  port->cnts.num_q_counters +
+
-					  port->cnts.num_cong_counters +
+	return rdma_alloc_hw_stats_struct(cnts->names,
-					  port->cnts.num_ext_ppcnt_counters,
+					  cnts->num_q_counters +
+					  cnts->num_cong_counters +
+					  cnts->num_ext_ppcnt_counters,
 					  RDMA_HW_STATS_DEFAULT_LIFESPAN);
 }
 
 static int mlx5_ib_query_q_counters(struct mlx5_core_dev *mdev,
-				    struct mlx5_ib_port *port,
+				    const struct mlx5_ib_counters *cnts,
 				    struct rdma_hw_stats *stats,
 				    u16 set_id)
 {
@@ -5501,8 +5544,8 @@ static int mlx5_ib_query_q_counters(struct mlx5_core_dev *mdev,
 	if (ret)
 		goto free;
 
-	for (i = 0; i < port->cnts.num_q_counters; i++) {
+	for (i = 0; i < cnts->num_q_counters; i++) {
-		val = *(__be32 *)(out + port->cnts.offsets[i]);
+		val = *(__be32 *)(out + cnts->offsets[i]);
 		stats->value[i] = (u64)be32_to_cpu(val);
 	}
 
@@ -5512,10 +5555,10 @@ free:
 }
 
 static int mlx5_ib_query_ext_ppcnt_counters(struct mlx5_ib_dev *dev,
-					  struct mlx5_ib_port *port,
+					    const struct mlx5_ib_counters *cnts,
-					  struct rdma_hw_stats *stats)
+					    struct rdma_hw_stats *stats)
 {
-	int offset = port->cnts.num_q_counters + port->cnts.num_cong_counters;
+	int offset = cnts->num_q_counters + cnts->num_cong_counters;
 	int sz = MLX5_ST_SZ_BYTES(ppcnt_reg);
 	int ret, i;
 	void *out;
@@ -5528,12 +5571,10 @@ static int mlx5_ib_query_ext_ppcnt_counters(struct mlx5_ib_dev *dev,
 	if (ret)
 		goto free;
 
-	for (i = 0; i < port->cnts.num_ext_ppcnt_counters; i++) {
+	for (i = 0; i < cnts->num_ext_ppcnt_counters; i++)
 		stats->value[i + offset] =
 			be64_to_cpup((__be64 *)(out +
-				    port->cnts.offsets[i + offset]));
+				    cnts->offsets[i + offset]));
-	}
-
 free:
 	kvfree(out);
 	return ret;
@@ -5544,7 +5585,7 @@ static int mlx5_ib_get_hw_stats(struct ib_device *ibdev,
 				u8 port_num, int index)
 {
 	struct mlx5_ib_dev *dev = to_mdev(ibdev);
-	struct mlx5_ib_port *port = &dev->port[port_num - 1];
+	const struct mlx5_ib_counters *cnts = get_counters(dev, port_num - 1);
 	struct mlx5_core_dev *mdev;
 	int ret, num_counters;
 	u8 mdev_port_num;
@@ -5552,18 +5593,17 @@ static int mlx5_ib_get_hw_stats(struct ib_device *ibdev,
 	if (!stats)
 		return -EINVAL;
 
-	num_counters = port->cnts.num_q_counters +
+	num_counters = cnts->num_q_counters +
-		       port->cnts.num_cong_counters +
+		       cnts->num_cong_counters +
-		       port->cnts.num_ext_ppcnt_counters;
+		       cnts->num_ext_ppcnt_counters;
 
 	/* q_counters are per IB device, query the master mdev */
-	ret = mlx5_ib_query_q_counters(dev->mdev, port, stats,
+	ret = mlx5_ib_query_q_counters(dev->mdev, cnts, stats, cnts->set_id);
-				       port->cnts.set_id);
 	if (ret)
 		return ret;
 
 	if (MLX5_CAP_PCAM_FEATURE(dev->mdev, rx_icrc_encapsulated_counter)) {
-		ret =  mlx5_ib_query_ext_ppcnt_counters(dev, port, stats);
+		ret =  mlx5_ib_query_ext_ppcnt_counters(dev, cnts, stats);
 		if (ret)
 			return ret;
 	}
@@ -5580,10 +5620,10 @@ static int mlx5_ib_get_hw_stats(struct ib_device *ibdev,
 		}
 		ret = mlx5_lag_query_cong_counters(dev->mdev,
 						   stats->value +
-						   port->cnts.num_q_counters,
+						   cnts->num_q_counters,
-						   port->cnts.num_cong_counters,
+						   cnts->num_cong_counters,
-						   port->cnts.offsets +
+						   cnts->offsets +
-						   port->cnts.num_q_counters);
+						   cnts->num_q_counters);
 
 		mlx5_ib_put_native_port_mdev(dev, port_num);
 		if (ret)
@@ -5598,20 +5638,22 @@ static struct rdma_hw_stats *
 mlx5_ib_counter_alloc_stats(struct rdma_counter *counter)
 {
 	struct mlx5_ib_dev *dev = to_mdev(counter->device);
-	struct mlx5_ib_port *port = &dev->port[counter->port - 1];
+	const struct mlx5_ib_counters *cnts =
+		get_counters(dev, counter->port - 1);
 
 	/* Q counters are in the beginning of all counters */
-	return rdma_alloc_hw_stats_struct(port->cnts.names,
+	return rdma_alloc_hw_stats_struct(cnts->names,
-					  port->cnts.num_q_counters,
+					  cnts->num_q_counters,
 					  RDMA_HW_STATS_DEFAULT_LIFESPAN);
 }
 
 static int mlx5_ib_counter_update_stats(struct rdma_counter *counter)
 {
 	struct mlx5_ib_dev *dev = to_mdev(counter->device);
-	struct mlx5_ib_port *port = &dev->port[counter->port - 1];
+	const struct mlx5_ib_counters *cnts =
+		get_counters(dev, counter->port - 1);
 
-	return mlx5_ib_query_q_counters(dev->mdev, port,
+	return mlx5_ib_query_q_counters(dev->mdev, cnts,
 					counter->stats, counter->id);
 }
 
@@ -5788,7 +5830,6 @@ static void init_delay_drop(struct mlx5_ib_dev *dev)
 		mlx5_ib_warn(dev, "Failed to init delay drop debugfs\n");
 }
 
-/* The mlx5_ib_multiport_mutex should be held when calling this function */
 static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
 				      struct mlx5_ib_multiport_info *mpi)
 {
@@ -5798,6 +5839,8 @@ static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
 	int err;
 	int i;
 
+	lockdep_assert_held(&mlx5_ib_multiport_mutex);
+
 	mlx5_ib_cleanup_cong_debugfs(ibdev, port_num);
 
 	spin_lock(&port->mp.mpi_lock);
@@ -5847,13 +5890,14 @@ static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
 	ibdev->port[port_num].roce.last_port_state = IB_PORT_DOWN;
 }
 
-/* The mlx5_ib_multiport_mutex should be held when calling this function */
 static bool mlx5_ib_bind_slave_port(struct mlx5_ib_dev *ibdev,
 				    struct mlx5_ib_multiport_info *mpi)
 {
 	u8 port_num = mlx5_core_native_port_num(mpi->mdev) - 1;
 	int err;
 
+	lockdep_assert_held(&mlx5_ib_multiport_mutex);
+
 	spin_lock(&ibdev->port[port_num].mp.mpi_lock);
 	if (ibdev->port[port_num].mp.mpi) {
 		mlx5_ib_dbg(ibdev, "port %d already affiliated.\n",
@@ -6882,7 +6926,7 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	dev->port = kcalloc(num_ports, sizeof(*dev->port),
 			     GFP_KERNEL);
 	if (!dev->port) {
-		ib_dealloc_device((struct ib_device *)dev);
+		ib_dealloc_device(&dev->ib_dev);
 		return NULL;
 	}
 
@@ -6909,6 +6953,7 @@ static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)
 			mlx5_ib_unbind_slave_port(mpi->ibdev, mpi);
 		list_del(&mpi->list);
 		mutex_unlock(&mlx5_ib_multiport_mutex);
+		kfree(mpi);
 		return;
 	}
 

drivers/infiniband/hw/mlx5/mlx5_ib.h

@@ -200,6 +200,7 @@ struct mlx5_ib_flow_db {
 	struct mlx5_ib_flow_prio	sniffer[MLX5_IB_NUM_SNIFFER_FTS];
 	struct mlx5_ib_flow_prio	egress[MLX5_IB_NUM_EGRESS_FTS];
 	struct mlx5_ib_flow_prio	fdb;
+	struct mlx5_ib_flow_prio	rdma_rx[MLX5_IB_NUM_FLOW_FT];
 	struct mlx5_flow_table		*lag_demux_ft;
 	/* Protect flow steering bypass flow tables
 	 * when add/del flow rules.
@@ -1476,6 +1477,7 @@ int bfregn_to_uar_index(struct mlx5_ib_dev *dev,
 			bool dyn_bfreg);
 
 int mlx5_ib_qp_set_counter(struct ib_qp *qp, struct rdma_counter *counter);
+u16 mlx5_ib_get_counters_id(struct mlx5_ib_dev *dev, u8 port_num);
 
 static inline bool mlx5_ib_can_use_umr(struct mlx5_ib_dev *dev,
 				       bool do_modify_atomic)

drivers/infiniband/hw/mlx5/odp.c

@@ -982,17 +982,6 @@ static int pagefault_data_segments(struct mlx5_ib_dev *dev,
 	return ret < 0 ? ret : npages;
 }
 
-static const u32 mlx5_ib_odp_opcode_cap[] = {
-	[MLX5_OPCODE_SEND]	       = IB_ODP_SUPPORT_SEND,
-	[MLX5_OPCODE_SEND_IMM]	       = IB_ODP_SUPPORT_SEND,
-	[MLX5_OPCODE_SEND_INVAL]       = IB_ODP_SUPPORT_SEND,
-	[MLX5_OPCODE_RDMA_WRITE]       = IB_ODP_SUPPORT_WRITE,
-	[MLX5_OPCODE_RDMA_WRITE_IMM]   = IB_ODP_SUPPORT_WRITE,
-	[MLX5_OPCODE_RDMA_READ]	       = IB_ODP_SUPPORT_READ,
-	[MLX5_OPCODE_ATOMIC_CS]	       = IB_ODP_SUPPORT_ATOMIC,
-	[MLX5_OPCODE_ATOMIC_FA]	       = IB_ODP_SUPPORT_ATOMIC,
-};
-
 /*
  * Parse initiator WQE. Advances the wqe pointer to point at the
  * scatter-gather list, and set wqe_end to the end of the WQE.
@@ -1003,12 +992,8 @@ static int mlx5_ib_mr_initiator_pfault_handler(
 {
 	struct mlx5_wqe_ctrl_seg *ctrl = *wqe;
 	u16 wqe_index = pfault->wqe.wqe_index;
-	u32 transport_caps;
 	struct mlx5_base_av *av;
 	unsigned ds, opcode;
-#if defined(DEBUG)
-	u32 ctrl_wqe_index, ctrl_qpn;
-#endif
 	u32 qpn = qp->trans_qp.base.mqp.qpn;
 
 	ds = be32_to_cpu(ctrl->qpn_ds) & MLX5_WQE_CTRL_DS_MASK;
@@ -1024,58 +1009,17 @@ static int mlx5_ib_mr_initiator_pfault_handler(
 		return -EFAULT;
 	}
 
-#if defined(DEBUG)
-	ctrl_wqe_index = (be32_to_cpu(ctrl->opmod_idx_opcode) &
-			MLX5_WQE_CTRL_WQE_INDEX_MASK) >>
-			MLX5_WQE_CTRL_WQE_INDEX_SHIFT;
-	if (wqe_index != ctrl_wqe_index) {
-		mlx5_ib_err(dev, "Got WQE with invalid wqe_index. wqe_index=0x%x, qpn=0x%x ctrl->wqe_index=0x%x\n",
-			    wqe_index, qpn,
-			    ctrl_wqe_index);
-		return -EFAULT;
-	}
-
-	ctrl_qpn = (be32_to_cpu(ctrl->qpn_ds) & MLX5_WQE_CTRL_QPN_MASK) >>
-		MLX5_WQE_CTRL_QPN_SHIFT;
-	if (qpn != ctrl_qpn) {
-		mlx5_ib_err(dev, "Got WQE with incorrect QP number. wqe_index=0x%x, qpn=0x%x ctrl->qpn=0x%x\n",
-			    wqe_index, qpn,
-			    ctrl_qpn);
-		return -EFAULT;
-	}
-#endif /* DEBUG */
-
 	*wqe_end = *wqe + ds * MLX5_WQE_DS_UNITS;
 	*wqe += sizeof(*ctrl);
 
 	opcode = be32_to_cpu(ctrl->opmod_idx_opcode) &
 		 MLX5_WQE_CTRL_OPCODE_MASK;
 
-	switch (qp->ibqp.qp_type) {
+	if (qp->ibqp.qp_type == IB_QPT_XRC_INI)
-	case IB_QPT_XRC_INI:
 		*wqe += sizeof(struct mlx5_wqe_xrc_seg);
-		transport_caps = dev->odp_caps.per_transport_caps.xrc_odp_caps;
-		break;
-	case IB_QPT_RC:
-		transport_caps = dev->odp_caps.per_transport_caps.rc_odp_caps;
-		break;
-	case IB_QPT_UD:
-		transport_caps = dev->odp_caps.per_transport_caps.ud_odp_caps;
-		break;
-	default:
-		mlx5_ib_err(dev, "ODP fault on QP of an unsupported transport 0x%x\n",
-			    qp->ibqp.qp_type);
-		return -EFAULT;
-	}
 
-	if (unlikely(opcode >= ARRAY_SIZE(mlx5_ib_odp_opcode_cap) ||
+	if (qp->ibqp.qp_type == IB_QPT_UD ||
-		     !(transport_caps & mlx5_ib_odp_opcode_cap[opcode]))) {
+	    qp->qp_sub_type == MLX5_IB_QPT_DCI) {
-		mlx5_ib_err(dev, "ODP fault on QP of an unsupported opcode 0x%x\n",
-			    opcode);
-		return -EFAULT;
-	}
-
-	if (qp->ibqp.qp_type == IB_QPT_UD) {
 		av = *wqe;
 		if (av->dqp_dct & cpu_to_be32(MLX5_EXTENDED_UD_AV))
 			*wqe += sizeof(struct mlx5_av);
@@ -1138,19 +1082,6 @@ static int mlx5_ib_mr_responder_pfault_handler_rq(struct mlx5_ib_dev *dev,
 		return -EFAULT;
 	}
 
-	switch (qp->ibqp.qp_type) {
-	case IB_QPT_RC:
-		if (!(dev->odp_caps.per_transport_caps.rc_odp_caps &
-		      IB_ODP_SUPPORT_RECV))
-			goto invalid_transport_or_opcode;
-		break;
-	default:
-invalid_transport_or_opcode:
-		mlx5_ib_err(dev, "ODP fault on QP of an unsupported transport. transport: 0x%x\n",
-			    qp->ibqp.qp_type);
-		return -EFAULT;
-	}
-
 	*wqe_end = wqe + wqe_size;
 
 	return 0;
@@ -1200,7 +1131,7 @@ static void mlx5_ib_mr_wqe_pfault_handler(struct mlx5_ib_dev *dev,
 {
 	bool sq = pfault->type & MLX5_PFAULT_REQUESTOR;
 	u16 wqe_index = pfault->wqe.wqe_index;
-	void *wqe = NULL, *wqe_end = NULL;
+	void *wqe, *wqe_start = NULL, *wqe_end = NULL;
 	u32 bytes_mapped, total_wqe_bytes;
 	struct mlx5_core_rsc_common *res;
 	int resume_with_error = 1;
@@ -1221,12 +1152,13 @@ static void mlx5_ib_mr_wqe_pfault_handler(struct mlx5_ib_dev *dev,
 		goto resolve_page_fault;
 	}
 
-	wqe = (void *)__get_free_page(GFP_KERNEL);
+	wqe_start = (void *)__get_free_page(GFP_KERNEL);
-	if (!wqe) {
+	if (!wqe_start) {
 		mlx5_ib_err(dev, "Error allocating memory for IO page fault handling.\n");
 		goto resolve_page_fault;
 	}
 
+	wqe = wqe_start;
 	qp = (res->res == MLX5_RES_QP) ? res_to_qp(res) : NULL;
 	if (qp && sq) {
 		ret = mlx5_ib_read_user_wqe_sq(qp, wqe_index, wqe, PAGE_SIZE,
@@ -1281,7 +1213,7 @@ resolve_page_fault:
 		    pfault->wqe.wq_num, resume_with_error,
 		    pfault->type);
 	mlx5_core_res_put(res);
-	free_page((unsigned long)wqe);
+	free_page((unsigned long)wqe_start);
 }
 
 static int pages_in_range(u64 address, u32 length)

drivers/infiniband/hw/mlx5/qp.c

@@ -3386,19 +3386,16 @@ static int __mlx5_ib_qp_set_counter(struct ib_qp *qp,
 	struct mlx5_ib_dev *dev = to_mdev(qp->device);
 	struct mlx5_ib_qp *mqp = to_mqp(qp);
 	struct mlx5_qp_context context = {};
-	struct mlx5_ib_port *mibport = NULL;
 	struct mlx5_ib_qp_base *base;
 	u32 set_id;
 
 	if (!MLX5_CAP_GEN(dev->mdev, rts2rts_qp_counters_set_id))
 		return 0;
 
-	if (counter) {
+	if (counter)
 		set_id = counter->id;
-	} else {
+	else
-		mibport = &dev->port[mqp->port - 1];
+		set_id = mlx5_ib_get_counters_id(dev, mqp->port - 1);
-		set_id = mibport->cnts.set_id;
-	}
 
 	base = &mqp->trans_qp.base;
 	context.qp_counter_set_usr_page &= cpu_to_be32(0xffffff);
@@ -3459,7 +3456,6 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 	struct mlx5_ib_cq *send_cq, *recv_cq;
 	struct mlx5_qp_context *context;
 	struct mlx5_ib_pd *pd;
-	struct mlx5_ib_port *mibport = NULL;
 	enum mlx5_qp_state mlx5_cur, mlx5_new;
 	enum mlx5_qp_optpar optpar;
 	u32 set_id = 0;
@@ -3624,11 +3620,10 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp,
 		if (qp->flags & MLX5_IB_QP_UNDERLAY)
 			port_num = 0;
 
-		mibport = &dev->port[port_num];
 		if (ibqp->counter)
 			set_id = ibqp->counter->id;
 		else
-			set_id = mibport->cnts.set_id;
+			set_id = mlx5_ib_get_counters_id(dev, port_num);
 		context->qp_counter_set_usr_page |=
 			cpu_to_be32(set_id << 24);
 	}
@@ -3817,6 +3812,8 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 
 	dctc = MLX5_ADDR_OF(create_dct_in, qp->dct.in, dct_context_entry);
 	if (cur_state == IB_QPS_RESET && new_state == IB_QPS_INIT) {
+		u16 set_id;
+
 		required |= IB_QP_ACCESS_FLAGS | IB_QP_PKEY_INDEX | IB_QP_PORT;
 		if (!is_valid_mask(attr_mask, required, 0))
 			return -EINVAL;
@@ -3843,7 +3840,9 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 		}
 		MLX5_SET(dctc, dctc, pkey_index, attr->pkey_index);
 		MLX5_SET(dctc, dctc, port, attr->port_num);
-		MLX5_SET(dctc, dctc, counter_set_id, dev->port[attr->port_num - 1].cnts.set_id);
+
+		set_id = mlx5_ib_get_counters_id(dev, attr->port_num - 1);
+		MLX5_SET(dctc, dctc, counter_set_id, set_id);
 
 	} else if (cur_state == IB_QPS_INIT && new_state == IB_QPS_RTR) {
 		struct mlx5_ib_modify_qp_resp resp = {};
@@ -6345,11 +6344,13 @@ int mlx5_ib_modify_wq(struct ib_wq *wq, struct ib_wq_attr *wq_attr,
 	}
 
 	if (curr_wq_state == IB_WQS_RESET && wq_state == IB_WQS_RDY) {
+		u16 set_id;
+
+		set_id = mlx5_ib_get_counters_id(dev, 0);
 		if (MLX5_CAP_GEN(dev->mdev, modify_rq_counter_set_id)) {
 			MLX5_SET64(modify_rq_in, in, modify_bitmask,
 				   MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_RQ_COUNTER_SET_ID);
-			MLX5_SET(rqc, rqc, counter_set_id,
+			MLX5_SET(rqc, rqc, counter_set_id, set_id);
-				 dev->port->cnts.set_id);
 		} else
 			dev_info_once(
 				&dev->ib_dev.dev,

drivers/infiniband/hw/ocrdma/ocrdma_verbs.c

@@ -163,10 +163,10 @@ int ocrdma_query_port(struct ib_device *ibdev,
 	netdev = dev->nic_info.netdev;
 	if (netif_running(netdev) && netif_oper_up(netdev)) {
 		port_state = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	} else {
 		port_state = IB_PORT_DOWN;
-		props->phys_state = 3;
+		props->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 	}
 	props->max_mtu = IB_MTU_4096;
 	props->active_mtu = iboe_get_mtu(netdev->mtu);

drivers/infiniband/hw/qedr/main.c

@@ -826,7 +826,7 @@ static int qedr_init_hw(struct qedr_dev *dev)
 	if (rc)
 		goto out;
 
-	dev->db_addr = (void __iomem *)(uintptr_t)out_params.dpi_addr;
+	dev->db_addr = out_params.dpi_addr;
 	dev->db_phys_addr = out_params.dpi_phys_addr;
 	dev->db_size = out_params.dpi_size;
 	dev->dpi = out_params.dpi;

drivers/infiniband/hw/qedr/qedr.h

@@ -229,7 +229,7 @@ struct qedr_ucontext {
 	struct ib_ucontext ibucontext;
 	struct qedr_dev *dev;
 	struct qedr_pd *pd;
-	u64 dpi_addr;
+	void __iomem *dpi_addr;
 	u64 dpi_phys_addr;
 	u32 dpi_size;
 	u16 dpi;

drivers/infiniband/hw/qedr/verbs.c

@@ -221,10 +221,10 @@ int qedr_query_port(struct ib_device *ibdev, u8 port, struct ib_port_attr *attr)
 	/* *attr being zeroed by the caller, avoid zeroing it here */
 	if (rdma_port->port_state == QED_RDMA_PORT_UP) {
 		attr->state = IB_PORT_ACTIVE;
-		attr->phys_state = 5;
+		attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	} else {
 		attr->state = IB_PORT_DOWN;
-		attr->phys_state = 3;
+		attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 	}
 	attr->max_mtu = IB_MTU_4096;
 	attr->active_mtu = iboe_get_mtu(dev->ndev->mtu);
@@ -2451,7 +2451,6 @@ int qedr_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
 	struct qedr_dev *dev = qp->dev;
 	struct ib_qp_attr attr;
 	int attr_mask = 0;
-	int rc = 0;
 
 	DP_DEBUG(dev, QEDR_MSG_QP, "destroy qp: destroying %p, qp type=%d\n",
 		 qp, qp->qp_type);
@@ -2496,7 +2495,7 @@ int qedr_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata)
 		xa_erase_irq(&dev->qps, qp->qp_id);
 		kfree(qp);
 	}
-	return rc;
+	return 0;
 }
 
 int qedr_create_ah(struct ib_ah *ibah, struct rdma_ah_attr *attr, u32 flags,

drivers/infiniband/hw/qib/qib_file_ops.c

@@ -1789,7 +1789,6 @@ static void unlock_expected_tids(struct qib_ctxtdata *rcd)
 
 static int qib_close(struct inode *in, struct file *fp)
 {
-	int ret = 0;
 	struct qib_filedata *fd;
 	struct qib_ctxtdata *rcd;
 	struct qib_devdata *dd;
@@ -1873,7 +1872,7 @@ static int qib_close(struct inode *in, struct file *fp)
 
 bail:
 	kfree(fd);
-	return ret;
+	return 0;
 }
 
 static int qib_ctxt_info(struct file *fp, struct qib_ctxt_info __user *uinfo)

drivers/infiniband/hw/qib/qib_rc.c

@@ -313,11 +313,8 @@ int qib_make_rc_req(struct rvt_qp *qp, unsigned long *flags)
 		case IB_WR_SEND:
 		case IB_WR_SEND_WITH_IMM:
 			/* If no credit, return. */
-			if (!(qp->s_flags & RVT_S_UNLIMITED_CREDIT) &&
+			if (!rvt_rc_credit_avail(qp, wqe))
-			    rvt_cmp_msn(wqe->ssn, qp->s_lsn + 1) > 0) {
-				qp->s_flags |= RVT_S_WAIT_SSN_CREDIT;
 				goto bail;
-			}
 			if (len > pmtu) {
 				qp->s_state = OP(SEND_FIRST);
 				len = pmtu;
@@ -344,11 +341,8 @@ int qib_make_rc_req(struct rvt_qp *qp, unsigned long *flags)
 			goto no_flow_control;
 		case IB_WR_RDMA_WRITE_WITH_IMM:
 			/* If no credit, return. */
-			if (!(qp->s_flags & RVT_S_UNLIMITED_CREDIT) &&
+			if (!rvt_rc_credit_avail(qp, wqe))
-			    rvt_cmp_msn(wqe->ssn, qp->s_lsn + 1) > 0) {
-				qp->s_flags |= RVT_S_WAIT_SSN_CREDIT;
 				goto bail;
-			}
 no_flow_control:
 			ohdr->u.rc.reth.vaddr =
 				cpu_to_be64(wqe->rdma_wr.remote_addr);

drivers/infiniband/hw/qib/qib_sysfs.c

@@ -436,6 +436,7 @@ QIB_DIAGC_ATTR(dmawait);
 QIB_DIAGC_ATTR(unaligned);
 QIB_DIAGC_ATTR(rc_dupreq);
 QIB_DIAGC_ATTR(rc_seqnak);
+QIB_DIAGC_ATTR(rc_crwaits);
 
 static struct attribute *diagc_default_attributes[] = {
 	&qib_diagc_attr_rc_resends.attr,
@@ -453,6 +454,7 @@ static struct attribute *diagc_default_attributes[] = {
 	&qib_diagc_attr_unaligned.attr,
 	&qib_diagc_attr_rc_dupreq.attr,
 	&qib_diagc_attr_rc_seqnak.attr,
+	&qib_diagc_attr_rc_crwaits.attr,
 	NULL
 };
 

drivers/infiniband/hw/usnic/usnic_ib_main.c

@@ -89,9 +89,15 @@ static void usnic_ib_dump_vf(struct usnic_ib_vf *vf, char *buf, int buf_sz)
 
 void usnic_ib_log_vf(struct usnic_ib_vf *vf)
 {
-	char buf[1000];
+	char *buf = kzalloc(1000, GFP_KERNEL);
-	usnic_ib_dump_vf(vf, buf, sizeof(buf));
+
+	if (!buf)
+		return;
+
+	usnic_ib_dump_vf(vf, buf, 1000);
 	usnic_dbg("%s\n", buf);
+
+	kfree(buf);
 }
 
 /* Start of netdev section */

drivers/infiniband/hw/usnic/usnic_ib_verbs.c

@@ -194,7 +194,7 @@ find_free_vf_and_create_qp_grp(struct usnic_ib_dev *us_ibdev,
 			return ERR_CAST(dev_list);
 		for (i = 0; dev_list[i]; i++) {
 			dev = dev_list[i];
-			vf = pci_get_drvdata(to_pci_dev(dev));
+			vf = dev_get_drvdata(dev);
 			spin_lock(&vf->lock);
 			vnic = vf->vnic;
 			if (!usnic_vnic_check_room(vnic, res_spec)) {
@@ -356,13 +356,14 @@ int usnic_ib_query_port(struct ib_device *ibdev, u8 port,
 
 	if (!us_ibdev->ufdev->link_up) {
 		props->state = IB_PORT_DOWN;
-		props->phys_state = 3;
+		props->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 	} else if (!us_ibdev->ufdev->inaddr) {
 		props->state = IB_PORT_INIT;
-		props->phys_state = 4;
+		props->phys_state =
+			IB_PORT_PHYS_STATE_PORT_CONFIGURATION_TRAINING;
 	} else {
 		props->state = IB_PORT_ACTIVE;
-		props->phys_state = 5;
+		props->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	}
 
 	props->port_cap_flags = 0;

drivers/infiniband/sw/rxe/rxe.h

@@ -65,10 +65,6 @@
  */
 #define RXE_UVERBS_ABI_VERSION		2
 
-#define RDMA_LINK_PHYS_STATE_LINK_UP	(5)
-#define RDMA_LINK_PHYS_STATE_DISABLED	(3)
-#define RDMA_LINK_PHYS_STATE_POLLING	(2)
-
 #define RXE_ROCE_V2_SPORT		(0xc000)
 
 static inline u32 rxe_crc32(struct rxe_dev *rxe,

drivers/infiniband/sw/rxe/rxe_param.h

@@ -154,7 +154,7 @@ enum rxe_port_param {
 	RXE_PORT_ACTIVE_WIDTH		= IB_WIDTH_1X,
 	RXE_PORT_ACTIVE_SPEED		= 1,
 	RXE_PORT_PKEY_TBL_LEN		= 64,
-	RXE_PORT_PHYS_STATE		= 2,
+	RXE_PORT_PHYS_STATE		= IB_PORT_PHYS_STATE_POLLING,
 	RXE_PORT_SUBNET_PREFIX		= 0xfe80000000000000ULL,
 };
 

drivers/infiniband/sw/rxe/rxe_verbs.c

@@ -69,11 +69,11 @@ static int rxe_query_port(struct ib_device *dev,
 			      &attr->active_width);
 
 	if (attr->state == IB_PORT_ACTIVE)
-		attr->phys_state = RDMA_LINK_PHYS_STATE_LINK_UP;
+		attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP;
 	else if (dev_get_flags(rxe->ndev) & IFF_UP)
-		attr->phys_state = RDMA_LINK_PHYS_STATE_POLLING;
+		attr->phys_state = IB_PORT_PHYS_STATE_POLLING;
 	else
-		attr->phys_state = RDMA_LINK_PHYS_STATE_DISABLED;
+		attr->phys_state = IB_PORT_PHYS_STATE_DISABLED;
 
 	mutex_unlock(&rxe->usdev_lock);
 

drivers/infiniband/sw/siw/siw_qp_tx.c

@@ -76,16 +76,15 @@ static int siw_try_1seg(struct siw_iwarp_tx *c_tx, void *paddr)
 			if (unlikely(!p))
 				return -EFAULT;
 
-			buffer = kmap_atomic(p);
+			buffer = kmap(p);
 
 			if (likely(PAGE_SIZE - off >= bytes)) {
 				memcpy(paddr, buffer + off, bytes);
-				kunmap_atomic(buffer);
 			} else {
 				unsigned long part = bytes - (PAGE_SIZE - off);
 
 				memcpy(paddr, buffer + off, part);
-				kunmap_atomic(buffer);
+				kunmap(p);
 
 				if (!mem->is_pbl)
 					p = siw_get_upage(mem->umem,
@@ -97,11 +96,10 @@ static int siw_try_1seg(struct siw_iwarp_tx *c_tx, void *paddr)
 				if (unlikely(!p))
 					return -EFAULT;
 
-				buffer = kmap_atomic(p);
+				buffer = kmap(p);
-				memcpy(paddr + part, buffer,
+				memcpy(paddr + part, buffer, bytes - part);
-				       bytes - part);
-				kunmap_atomic(buffer);
 			}
+			kunmap(p);
 		}
 	}
 	return (int)bytes;
@@ -518,11 +516,12 @@ static int siw_tx_hdt(struct siw_iwarp_tx *c_tx, struct socket *s)
 							c_tx->mpa_crc_hd,
 							iov[seg].iov_base,
 							plen);
-				} else if (do_crc)
+				} else if (do_crc) {
-					crypto_shash_update(
+					crypto_shash_update(c_tx->mpa_crc_hd,
-						c_tx->mpa_crc_hd,
+							    kmap(p) + fp_off,
-						page_address(p) + fp_off,
+							    plen);
-						plen);
+					kunmap(p);
+				}
 			} else {
 				u64 va = sge->laddr + sge_off;
 

drivers/infiniband/sw/siw/siw_verbs.c

@@ -206,7 +206,8 @@ int siw_query_port(struct ib_device *base_dev, u8 port,
 	attr->gid_tbl_len = 1;
 	attr->max_msg_sz = -1;
 	attr->max_mtu = ib_mtu_int_to_enum(sdev->netdev->mtu);
-	attr->phys_state = sdev->state == IB_PORT_ACTIVE ? 5 : 3;
+	attr->phys_state = sdev->state == IB_PORT_ACTIVE ?
+		IB_PORT_PHYS_STATE_LINK_UP : IB_PORT_PHYS_STATE_DISABLED;
 	attr->pkey_tbl_len = 1;
 	attr->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_DEVICE_MGMT_SUP;
 	attr->state = sdev->state;

drivers/infiniband/ulp/iser/iscsi_iser.h

@@ -102,9 +102,10 @@
 
 /* Default support is 512KB I/O size */
 #define ISER_DEF_MAX_SECTORS		1024
-#define ISCSI_ISER_DEF_SG_TABLESIZE	((ISER_DEF_MAX_SECTORS * 512) >> SHIFT_4K)
+#define ISCSI_ISER_DEF_SG_TABLESIZE                                            \
-/* Maximum support is 8MB I/O size */
+	((ISER_DEF_MAX_SECTORS * SECTOR_SIZE) >> SHIFT_4K)
-#define ISCSI_ISER_MAX_SG_TABLESIZE	((16384 * 512) >> SHIFT_4K)
+/* Maximum support is 16MB I/O size */
+#define ISCSI_ISER_MAX_SG_TABLESIZE	((32768 * SECTOR_SIZE) >> SHIFT_4K)
 
 #define ISER_DEF_XMIT_CMDS_DEFAULT		512
 #if ISCSI_DEF_XMIT_CMDS_MAX > ISER_DEF_XMIT_CMDS_DEFAULT

drivers/infiniband/ulp/srpt/ib_srpt.c

@@ -1767,8 +1767,8 @@ static int srpt_create_ch_ib(struct srpt_rdma_ch *ch)
 		goto out;
 
 retry:
-	ch->cq = ib_alloc_cq(sdev->device, ch, ch->rq_size + sq_size,
+	ch->cq = ib_alloc_cq_any(sdev->device, ch, ch->rq_size + sq_size,
-			0 /* XXX: spread CQs */, IB_POLL_WORKQUEUE);
+				 IB_POLL_WORKQUEUE);
 	if (IS_ERR(ch->cq)) {
 		ret = PTR_ERR(ch->cq);
 		pr_err("failed to create CQ cqe= %d ret= %d\n",

drivers/net/ethernet/mellanox/mlx5/core/rl.c

@@ -188,8 +188,7 @@ int mlx5_rl_add_rate(struct mlx5_core_dev *dev, u16 *index,
 		/* new rate limit */
 		err = mlx5_set_pp_rate_limit_cmd(dev, entry->index, rl);
 		if (err) {
-			mlx5_core_err(dev, "Failed configuring rate limit(err %d): \
+			mlx5_core_err(dev, "Failed configuring rate limit(err %d): rate %u, max_burst_sz %u, typical_pkt_sz %u\n",
-				      rate %u, max_burst_sz %u, typical_pkt_sz %u\n",
 				      err, rl->rate, rl->max_burst_sz,
 				      rl->typical_pkt_sz);
 			goto out;
@@ -218,8 +217,7 @@ void mlx5_rl_remove_rate(struct mlx5_core_dev *dev, struct mlx5_rate_limit *rl)
 	mutex_lock(&table->rl_lock);
 	entry = find_rl_entry(table, rl);
 	if (!entry || !entry->refcount) {
-		mlx5_core_warn(dev, "Rate %u, max_burst_sz %u typical_pkt_sz %u \
+		mlx5_core_warn(dev, "Rate %u, max_burst_sz %u typical_pkt_sz %u are not configured\n",
-			       are not configured\n",
 			       rl->rate, rl->max_burst_sz, rl->typical_pkt_sz);
 		goto out;
 	}

drivers/net/ethernet/qlogic/qed/qed_rdma.c

@@ -798,9 +798,8 @@ static int qed_rdma_add_user(void *rdma_cxt,
 	/* Calculate the corresponding DPI address */
 	dpi_start_offset = p_hwfn->dpi_start_offset;
 
-	out_params->dpi_addr = (u64)((u8 __iomem *)p_hwfn->doorbells +
+	out_params->dpi_addr = p_hwfn->doorbells + dpi_start_offset +
-				     dpi_start_offset +
+			       out_params->dpi * p_hwfn->dpi_size;
-				     ((out_params->dpi) * p_hwfn->dpi_size));
 
 	out_params->dpi_phys_addr = p_hwfn->db_phys_addr +
 				    dpi_start_offset +

fs/cifs/smbdirect.c

@@ -1654,15 +1654,17 @@ static struct smbd_connection *_smbd_get_connection(
 
 	info->send_cq = NULL;
 	info->recv_cq = NULL;
-	info->send_cq = ib_alloc_cq(info->id->device, info,
+	info->send_cq =
-			info->send_credit_target, 0, IB_POLL_SOFTIRQ);
+		ib_alloc_cq_any(info->id->device, info,
+				info->send_credit_target, IB_POLL_SOFTIRQ);
 	if (IS_ERR(info->send_cq)) {
 		info->send_cq = NULL;
 		goto alloc_cq_failed;
 	}
 
-	info->recv_cq = ib_alloc_cq(info->id->device, info,
+	info->recv_cq =
-			info->receive_credit_max, 0, IB_POLL_SOFTIRQ);
+		ib_alloc_cq_any(info->id->device, info,
+				info->receive_credit_max, IB_POLL_SOFTIRQ);
 	if (IS_ERR(info->recv_cq)) {
 		info->recv_cq = NULL;
 		goto alloc_cq_failed;

include/Kbuild

@@ -881,12 +881,6 @@ header-test-			+= net/xdp.h
 header-test-			+= net/xdp_priv.h
 header-test-			+= pcmcia/cistpl.h
 header-test-			+= pcmcia/ds.h
-header-test-			+= rdma/ib.h
-header-test-			+= rdma/iw_portmap.h
-header-test-			+= rdma/opa_port_info.h
-header-test-			+= rdma/rdmavt_cq.h
-header-test-			+= rdma/restrack.h
-header-test-			+= rdma/signature.h
 header-test-			+= rdma/tid_rdma_defs.h
 header-test-			+= scsi/fc/fc_encaps.h
 header-test-			+= scsi/fc/fc_fc2.h

include/linux/mlx5/device.h

@@ -328,6 +328,7 @@ enum mlx5_event {
 	MLX5_EVENT_TYPE_GPIO_EVENT	   = 0x15,
 	MLX5_EVENT_TYPE_PORT_MODULE_EVENT  = 0x16,
 	MLX5_EVENT_TYPE_TEMP_WARN_EVENT    = 0x17,
+	MLX5_EVENT_TYPE_XRQ_ERROR	   = 0x18,
 	MLX5_EVENT_TYPE_REMOTE_CONFIG	   = 0x19,
 	MLX5_EVENT_TYPE_GENERAL_EVENT	   = 0x22,
 	MLX5_EVENT_TYPE_MONITOR_COUNTER    = 0x24,
@@ -345,6 +346,7 @@ enum mlx5_event {
 	MLX5_EVENT_TYPE_ESW_FUNCTIONS_CHANGED = 0xe,
 
 	MLX5_EVENT_TYPE_DCT_DRAINED        = 0x1c,
+	MLX5_EVENT_TYPE_DCT_KEY_VIOLATION  = 0x1d,
 
 	MLX5_EVENT_TYPE_FPGA_ERROR         = 0x20,
 	MLX5_EVENT_TYPE_FPGA_QP_ERROR      = 0x21,
@@ -584,6 +586,12 @@ struct mlx5_eqe_cq_err {
 	u8	syndrome;
 };
 
+struct mlx5_eqe_xrq_err {
+	__be32	reserved1[5];
+	__be32	type_xrqn;
+	__be32	reserved2;
+};
+
 struct mlx5_eqe_port_state {
 	u8	reserved0[8];
 	u8	port;
@@ -698,6 +706,7 @@ union ev_data {
 	struct mlx5_eqe_pps		pps;
 	struct mlx5_eqe_dct             dct;
 	struct mlx5_eqe_temp_warning	temp_warning;
+	struct mlx5_eqe_xrq_err		xrq_err;
 } __packed;
 
 struct mlx5_eqe {

include/linux/qed/qed_rdma_if.h

@@ -225,7 +225,7 @@ struct qed_rdma_start_in_params {
 
 struct qed_rdma_add_user_out_params {
 	u16 dpi;
-	u64 dpi_addr;
+	void __iomem *dpi_addr;
 	u64 dpi_phys_addr;
 	u32 dpi_size;
 	u16 wid_count;

include/rdma/ib.h

@@ -36,6 +36,8 @@
 #include <linux/types.h>
 #include <linux/sched.h>
 #include <linux/cred.h>
+#include <linux/uaccess.h>
+#include <linux/fs.h>
 
 struct ib_addr {
 	union {

include/rdma/ib_verbs.h

@@ -98,15 +98,54 @@ void ibdev_info(const struct ib_device *ibdev, const char *format, ...);
 #if defined(CONFIG_DYNAMIC_DEBUG)
 #define ibdev_dbg(__dev, format, args...)                       \
 	dynamic_ibdev_dbg(__dev, format, ##args)
-#elif defined(DEBUG)
-#define ibdev_dbg(__dev, format, args...)                       \
-	ibdev_printk(KERN_DEBUG, __dev, format, ##args)
 #else
 __printf(2, 3) __cold
 static inline
 void ibdev_dbg(const struct ib_device *ibdev, const char *format, ...) {}
 #endif
 
+#define ibdev_level_ratelimited(ibdev_level, ibdev, fmt, ...)           \
+do {                                                                    \
+	static DEFINE_RATELIMIT_STATE(_rs,                              \
+				      DEFAULT_RATELIMIT_INTERVAL,       \
+				      DEFAULT_RATELIMIT_BURST);         \
+	if (__ratelimit(&_rs))                                          \
+		ibdev_level(ibdev, fmt, ##__VA_ARGS__);                 \
+} while (0)
+
+#define ibdev_emerg_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_emerg, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_alert_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_alert, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_crit_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_crit, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_err_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_err, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_warn_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_warn, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_notice_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_notice, ibdev, fmt, ##__VA_ARGS__)
+#define ibdev_info_ratelimited(ibdev, fmt, ...) \
+	ibdev_level_ratelimited(ibdev_info, ibdev, fmt, ##__VA_ARGS__)
+
+#if defined(CONFIG_DYNAMIC_DEBUG)
+/* descriptor check is first to prevent flooding with "callbacks suppressed" */
+#define ibdev_dbg_ratelimited(ibdev, fmt, ...)                          \
+do {                                                                    \
+	static DEFINE_RATELIMIT_STATE(_rs,                              \
+				      DEFAULT_RATELIMIT_INTERVAL,       \
+				      DEFAULT_RATELIMIT_BURST);         \
+	DEFINE_DYNAMIC_DEBUG_METADATA(descriptor, fmt);                 \
+	if (DYNAMIC_DEBUG_BRANCH(descriptor) && __ratelimit(&_rs))      \
+		__dynamic_ibdev_dbg(&descriptor, ibdev, fmt,            \
+				    ##__VA_ARGS__);                     \
+} while (0)
+#else
+__printf(2, 3) __cold
+static inline
+void ibdev_dbg_ratelimited(const struct ib_device *ibdev, const char *format, ...) {}
+#endif
+
 union ib_gid {
 	u8	raw[16];
 	struct {
@@ -451,6 +490,16 @@ enum ib_port_state {
 	IB_PORT_ACTIVE_DEFER	= 5
 };
 
+enum ib_port_phys_state {
+	IB_PORT_PHYS_STATE_SLEEP = 1,
+	IB_PORT_PHYS_STATE_POLLING = 2,
+	IB_PORT_PHYS_STATE_DISABLED = 3,
+	IB_PORT_PHYS_STATE_PORT_CONFIGURATION_TRAINING = 4,
+	IB_PORT_PHYS_STATE_LINK_UP = 5,
+	IB_PORT_PHYS_STATE_LINK_ERROR_RECOVERY = 6,
+	IB_PORT_PHYS_STATE_PHY_TEST = 7,
+};
+
 enum ib_port_width {
 	IB_WIDTH_1X	= 1,
 	IB_WIDTH_2X	= 16,
@@ -3710,6 +3759,25 @@ static inline struct ib_cq *ib_alloc_cq(struct ib_device *dev, void *private,
 				NULL);
 }
 
+struct ib_cq *__ib_alloc_cq_any(struct ib_device *dev, void *private,
+				int nr_cqe, enum ib_poll_context poll_ctx,
+				const char *caller);
+
+/**
+ * ib_alloc_cq_any: Allocate kernel CQ
+ * @dev: The IB device
+ * @private: Private data attached to the CQE
+ * @nr_cqe: Number of CQEs in the CQ
+ * @poll_ctx: Context used for polling the CQ
+ */
+static inline struct ib_cq *ib_alloc_cq_any(struct ib_device *dev,
+					    void *private, int nr_cqe,
+					    enum ib_poll_context poll_ctx)
+{
+	return __ib_alloc_cq_any(dev, private, nr_cqe, poll_ctx,
+				 KBUILD_MODNAME);
+}
+
 /**
  * ib_free_cq_user - Free kernel/user CQ
  * @cq: The CQ to free

include/rdma/iw_portmap.h

@@ -33,6 +33,9 @@
 #ifndef _IW_PORTMAP_H
 #define _IW_PORTMAP_H
 
+#include <linux/socket.h>
+#include <linux/netlink.h>
+
 #define IWPM_ULIBNAME_SIZE	32
 #define IWPM_DEVNAME_SIZE	32
 #define IWPM_IFNAME_SIZE	16

include/rdma/opa_port_info.h

@@ -33,6 +33,8 @@
 #if !defined(OPA_PORT_INFO_H)
 #define OPA_PORT_INFO_H
 
+#include <rdma/opa_smi.h>
+
 #define OPA_PORT_LINK_MODE_NOP	0		/* No change */
 #define OPA_PORT_LINK_MODE_OPA	4		/* Port mode is OPA */
 

include/rdma/rdma_netlink.h

@@ -76,28 +76,32 @@ int ibnl_put_attr(struct sk_buff *skb, struct nlmsghdr *nlh,
 
 /**
  * Send the supplied skb to a specific userspace PID.
+ * @net: Net namespace in which to send the skb
  * @skb: The netlink skb
  * @pid: Userspace netlink process ID
  * Returns 0 on success or a negative error code.
  */
-int rdma_nl_unicast(struct sk_buff *skb, u32 pid);
+int rdma_nl_unicast(struct net *net, struct sk_buff *skb, u32 pid);
 
 /**
  * Send, with wait/1 retry, the supplied skb to a specific userspace PID.
+ * @net: Net namespace in which to send the skb
  * @skb: The netlink skb
  * @pid: Userspace netlink process ID
  * Returns 0 on success or a negative error code.
  */
-int rdma_nl_unicast_wait(struct sk_buff *skb, __u32 pid);
+int rdma_nl_unicast_wait(struct net *net, struct sk_buff *skb, __u32 pid);
 
 /**
  * Send the supplied skb to a netlink group.
+ * @net: Net namespace in which to send the skb
  * @skb: The netlink skb
  * @group: Netlink group ID
  * @flags: allocation flags
  * Returns 0 on success or a negative error code.
  */
-int rdma_nl_multicast(struct sk_buff *skb, unsigned int group, gfp_t flags);
+int rdma_nl_multicast(struct net *net, struct sk_buff *skb,
+		      unsigned int group, gfp_t flags);
 
 /**
  * Check if there are any listeners to the netlink group

include/rdma/rdma_vt.h

@@ -116,6 +116,7 @@ struct rvt_ibport {
 	u64 n_unaligned;
 	u64 n_rc_dupreq;
 	u64 n_rc_seqnak;
+	u64 n_rc_crwaits;
 	u16 pkey_violations;
 	u16 qkey_violations;
 	u16 mkey_violations;

include/rdma/rdmavt_cq.h

@@ -53,6 +53,7 @@
 
 #include <linux/kthread.h>
 #include <rdma/ib_user_verbs.h>
+#include <rdma/ib_verbs.h>
 
 /*
  * Define an ib_cq_notify value that is not valid so we know when CQ

include/rdma/rdmavt_qp.h

@@ -973,6 +973,41 @@ static inline void rvt_free_rq(struct rvt_rq *rq)
 	rq->wq = NULL;
 }
 
+/**
+ * rvt_to_iport - Get the ibport pointer
+ * @qp: the qp pointer
+ *
+ * This function returns the ibport pointer from the qp pointer.
+ */
+static inline struct rvt_ibport *rvt_to_iport(struct rvt_qp *qp)
+{
+	struct rvt_dev_info *rdi = ib_to_rvt(qp->ibqp.device);
+
+	return rdi->ports[qp->port_num - 1];
+}
+
+/**
+ * rvt_rc_credit_avail - Check if there are enough RC credits for the request
+ * @qp: the qp
+ * @wqe: the request
+ *
+ * This function returns false when there are not enough credits for the given
+ * request and true otherwise.
+ */
+static inline bool rvt_rc_credit_avail(struct rvt_qp *qp, struct rvt_swqe *wqe)
+{
+	lockdep_assert_held(&qp->s_lock);
+	if (!(qp->s_flags & RVT_S_UNLIMITED_CREDIT) &&
+	    rvt_cmp_msn(wqe->ssn, qp->s_lsn + 1) > 0) {
+		struct rvt_ibport *rvp = rvt_to_iport(qp);
+
+		qp->s_flags |= RVT_S_WAIT_SSN_CREDIT;
+		rvp->n_rc_crwaits++;
+		return false;
+	}
+	return true;
+}
+
 struct rvt_qp_iter *rvt_qp_iter_init(struct rvt_dev_info *rdi,
 				     u64 v,
 				     void (*cb)(struct rvt_qp *qp, u64 v));

include/rdma/signature.h

@@ -6,6 +6,8 @@
 #ifndef _RDMA_SIGNATURE_H_
 #define _RDMA_SIGNATURE_H_
 
+#include <linux/types.h>
+
 enum ib_signature_prot_cap {
 	IB_PROT_T10DIF_TYPE_1 = 1,
 	IB_PROT_T10DIF_TYPE_2 = 1 << 1,

include/uapi/rdma/mlx5_user_ioctl_verbs.h

@@ -43,6 +43,7 @@ enum mlx5_ib_uapi_flow_table_type {
 	MLX5_IB_UAPI_FLOW_TABLE_TYPE_NIC_RX     = 0x0,
 	MLX5_IB_UAPI_FLOW_TABLE_TYPE_NIC_TX	= 0x1,
 	MLX5_IB_UAPI_FLOW_TABLE_TYPE_FDB	= 0x2,
+	MLX5_IB_UAPI_FLOW_TABLE_TYPE_RDMA_RX	= 0x3,
 };
 
 enum mlx5_ib_uapi_flow_action_packet_reformat_type {

net/9p/trans_rdma.c

@@ -685,9 +685,9 @@ rdma_create_trans(struct p9_client *client, const char *addr, char *args)
 		goto error;
 
 	/* Create the Completion Queue */
-	rdma->cq = ib_alloc_cq(rdma->cm_id->device, client,
+	rdma->cq = ib_alloc_cq_any(rdma->cm_id->device, client,
-			opts.sq_depth + opts.rq_depth + 1,
+				   opts.sq_depth + opts.rq_depth + 1,
-			0, IB_POLL_SOFTIRQ);
+				   IB_POLL_SOFTIRQ);
 	if (IS_ERR(rdma->cq))
 		goto error;
 

net/sunrpc/xprtrdma/svc_rdma_transport.c

@@ -454,14 +454,14 @@ static struct svc_xprt *svc_rdma_accept(struct svc_xprt *xprt)
 		dprintk("svcrdma: error creating PD for connect request\n");
 		goto errout;
 	}
-	newxprt->sc_sq_cq = ib_alloc_cq(dev, newxprt, newxprt->sc_sq_depth,
+	newxprt->sc_sq_cq = ib_alloc_cq_any(dev, newxprt, newxprt->sc_sq_depth,
-					0, IB_POLL_WORKQUEUE);
+					    IB_POLL_WORKQUEUE);
 	if (IS_ERR(newxprt->sc_sq_cq)) {
 		dprintk("svcrdma: error creating SQ CQ for connect request\n");
 		goto errout;
 	}
-	newxprt->sc_rq_cq = ib_alloc_cq(dev, newxprt, rq_depth,
+	newxprt->sc_rq_cq =
-					0, IB_POLL_WORKQUEUE);
+		ib_alloc_cq_any(dev, newxprt, rq_depth, IB_POLL_WORKQUEUE);
 	if (IS_ERR(newxprt->sc_rq_cq)) {
 		dprintk("svcrdma: error creating RQ CQ for connect request\n");
 		goto errout;

net/sunrpc/xprtrdma/verbs.c

@@ -521,18 +521,17 @@ int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
 	init_waitqueue_head(&ep->rep_connect_wait);
 	ep->rep_receive_count = 0;
 
-	sendcq = ib_alloc_cq(ia->ri_id->device, NULL,
+	sendcq = ib_alloc_cq_any(ia->ri_id->device, NULL,
-			     ep->rep_attr.cap.max_send_wr + 1,
+				 ep->rep_attr.cap.max_send_wr + 1,
-			     ia->ri_id->device->num_comp_vectors > 1 ? 1 : 0,
+				 IB_POLL_WORKQUEUE);
-			     IB_POLL_WORKQUEUE);
 	if (IS_ERR(sendcq)) {
 		rc = PTR_ERR(sendcq);
 		goto out1;
 	}
 
-	recvcq = ib_alloc_cq(ia->ri_id->device, NULL,
+	recvcq = ib_alloc_cq_any(ia->ri_id->device, NULL,
-			     ep->rep_attr.cap.max_recv_wr + 1,
+				 ep->rep_attr.cap.max_recv_wr + 1,
-			     0, IB_POLL_WORKQUEUE);
+				 IB_POLL_WORKQUEUE);
 	if (IS_ERR(recvcq)) {
 		rc = PTR_ERR(recvcq);
 		goto out2;