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

Pull rdma fixes from Doug Ledford:
 "For the most part this is just a minor -rc cycle for the rdma
  subsystem. Even given that this is all of the -rc patches since the
  merge window closed, it's still only about 25 patches:

   - Multiple i40iw, nes, iw_cxgb4, hfi1, qib, mlx4, mlx5 fixes

   - A few upper layer protocol fixes (IPoIB, iSER, SRP)

   - A modest number of core fixes"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (26 commits)
  RDMA/SA: Fix kernel panic in CMA request handler flow
  RDMA/umem: Fix missing mmap_sem in get umem ODP call
  RDMA/core: not to set page dirty bit if it's already set.
  RDMA/uverbs: Declare local function static and add brackets to sizeof
  RDMA/netlink: Reduce exposure of RDMA netlink functions
  RDMA/srp: Fix NULL deref at srp_destroy_qp()
  RDMA/IPoIB: Limit the ipoib_dev_uninit_default scope
  RDMA/IPoIB: Replace netdev_priv with ipoib_priv for ipoib_get_link_ksettings
  RDMA/qedr: add null check before pointer dereference
  RDMA/mlx5: set UMR wqe fence according to HCA cap
  net/mlx5: Define interface bits for fencing UMR wqe
  RDMA/mlx4: Fix MAD tunneling when SRIOV is enabled
  RDMA/qib,hfi1: Fix MR reference count leak on write with immediate
  RDMA/hfi1: Defer setting VL15 credits to link-up interrupt
  RDMA/hfi1: change PCI bar addr assignments to Linux API functions
  RDMA/hfi1: fix array termination by appending NULL to attr array
  RDMA/iw_cxgb4: fix the calculation of ipv6 header size
  RDMA/iw_cxgb4: calculate t4_eq_status_entries properly
  RDMA/iw_cxgb4: Avoid touch after free error in ARP failure handlers
  RDMA/nes: ACK MPA Reply frame
  ...

37 files changed, 194 insertions, 170 deletions

diff --git a/drivers/infiniband/core/cm.c b/drivers/infiniband/core/cm.c
index 1844770f3ae8..2b4d613a3474 100644
--- a/drivers/infiniband/core/cm.c
+++ b/drivers/infiniband/core/cm.c
@@ -1429,7 +1429,7 @@ static void cm_format_paths_from_req(struct cm_req_msg *req_msg,
 	primary_path->packet_life_time =
 		cm_req_get_primary_local_ack_timeout(req_msg);
 	primary_path->packet_life_time -= (primary_path->packet_life_time > 0);
-	sa_path_set_service_id(primary_path, req_msg->service_id);
+	primary_path->service_id = req_msg->service_id;
 
 	if (req_msg->alt_local_lid) {
 		alt_path->dgid = req_msg->alt_local_gid;
@@ -1452,7 +1452,7 @@ static void cm_format_paths_from_req(struct cm_req_msg *req_msg,
 		alt_path->packet_life_time =
 			cm_req_get_alt_local_ack_timeout(req_msg);
 		alt_path->packet_life_time -= (alt_path->packet_life_time > 0);
-		sa_path_set_service_id(alt_path, req_msg->service_id);
+		alt_path->service_id = req_msg->service_id;
 	}
 }
 
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 91b7a2fe5a55..31bb82d8ecd7 100644
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -1140,7 +1140,7 @@ static void cma_save_ib_info(struct sockaddr *src_addr,
 			ib->sib_pkey = path->pkey;
 			ib->sib_flowinfo = path->flow_label;
 			memcpy(&ib->sib_addr, &path->sgid, 16);
-			ib->sib_sid = sa_path_get_service_id(path);
+			ib->sib_sid = path->service_id;
 			ib->sib_scope_id = 0;
 		} else {
 			ib->sib_pkey = listen_ib->sib_pkey;
@@ -1274,8 +1274,7 @@ static int cma_save_req_info(const struct ib_cm_event *ib_event,
 		memcpy(&req->local_gid, &req_param->primary_path->sgid,
 		       sizeof(req->local_gid));
 		req->has_gid	= true;
-		req->service_id	=
-			sa_path_get_service_id(req_param->primary_path);
+		req->service_id = req_param->primary_path->service_id;
 		req->pkey	= be16_to_cpu(req_param->primary_path->pkey);
 		if (req->pkey != req_param->bth_pkey)
 			pr_warn_ratelimited("RDMA CMA: got different BTH P_Key (0x%x) and primary path P_Key (0x%x)\n"
@@ -1827,7 +1826,8 @@ static struct rdma_id_private *cma_new_conn_id(struct rdma_cm_id *listen_id,
 	struct rdma_route *rt;
 	const sa_family_t ss_family = listen_id->route.addr.src_addr.ss_family;
 	struct sa_path_rec *path = ib_event->param.req_rcvd.primary_path;
-	const __be64 service_id = sa_path_get_service_id(path);
+	const __be64 service_id =
+		ib_event->param.req_rcvd.primary_path->service_id;
 	int ret;
 
 	id = rdma_create_id(listen_id->route.addr.dev_addr.net,
@@ -2345,9 +2345,8 @@ static int cma_query_ib_route(struct rdma_id_private *id_priv, int timeout_ms,
 	path_rec.pkey = cpu_to_be16(ib_addr_get_pkey(dev_addr));
 	path_rec.numb_path = 1;
 	path_rec.reversible = 1;
-	sa_path_set_service_id(&path_rec,
-			       rdma_get_service_id(&id_priv->id,
-						   cma_dst_addr(id_priv)));
+	path_rec.service_id = rdma_get_service_id(&id_priv->id,
+						  cma_dst_addr(id_priv));
 
 	comp_mask = IB_SA_PATH_REC_DGID | IB_SA_PATH_REC_SGID |
 		    IB_SA_PATH_REC_PKEY | IB_SA_PATH_REC_NUMB_PATH |
diff --git a/drivers/infiniband/core/core_priv.h b/drivers/infiniband/core/core_priv.h
index cb7d372e4bdf..d92ab4eaa8f3 100644
--- a/drivers/infiniband/core/core_priv.h
+++ b/drivers/infiniband/core/core_priv.h
@@ -169,6 +169,16 @@ void ib_mad_cleanup(void);
 int ib_sa_init(void);
 void ib_sa_cleanup(void);
 
+int ibnl_init(void);
+void ibnl_cleanup(void);
+
+/**
+ * Check if there are any listeners to the netlink group
+ * @group: the netlink group ID
+ * Returns 0 on success or a negative for no listeners.
+ */
+int ibnl_chk_listeners(unsigned int group);
+
 int ib_nl_handle_resolve_resp(struct sk_buff *skb,
 			      struct netlink_callback *cb);
 int ib_nl_handle_set_timeout(struct sk_buff *skb,
diff --git a/drivers/infiniband/core/netlink.c b/drivers/infiniband/core/netlink.c
index b784055423c8..94931c474d41 100644
--- a/drivers/infiniband/core/netlink.c
+++ b/drivers/infiniband/core/netlink.c
@@ -37,6 +37,7 @@
 #include <net/net_namespace.h>
 #include <net/sock.h>
 #include <rdma/rdma_netlink.h>
+#include "core_priv.h"
 
 struct ibnl_client {
 	struct list_head		list;
@@ -55,7 +56,6 @@ int ibnl_chk_listeners(unsigned int group)
 		return -1;
 	return 0;
 }
-EXPORT_SYMBOL(ibnl_chk_listeners);
 
 int ibnl_add_client(int index, int nops,
 		    const struct ibnl_client_cbs cb_table[])
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index e335b09c022e..fb7aec4047c8 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -194,7 +194,7 @@ static u32 tid;
 	.field_name          = "sa_path_rec:" #field
 
 static const struct ib_field path_rec_table[] = {
-	{ PATH_REC_FIELD(ib.service_id),
+	{ PATH_REC_FIELD(service_id),
 	  .offset_words = 0,
 	  .offset_bits  = 0,
 	  .size_bits    = 64 },
@@ -296,7 +296,7 @@ static const struct ib_field path_rec_table[] = {
 	.field_name          = "sa_path_rec:" #field
 
 static const struct ib_field opa_path_rec_table[] = {
-	{ OPA_PATH_REC_FIELD(opa.service_id),
+	{ OPA_PATH_REC_FIELD(service_id),
 	  .offset_words = 0,
 	  .offset_bits  = 0,
 	  .size_bits    = 64 },
@@ -774,7 +774,7 @@ static void ib_nl_set_path_rec_attrs(struct sk_buff *skb,
 
 	/* Now build the attributes */
 	if (comp_mask & IB_SA_PATH_REC_SERVICE_ID) {
-		val64 = be64_to_cpu(sa_path_get_service_id(sa_rec));
+		val64 = be64_to_cpu(sa_rec->service_id);
 		nla_put(skb, RDMA_NLA_F_MANDATORY | LS_NLA_TYPE_SERVICE_ID,
 			sizeof(val64), &val64);
 	}
diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
index 3dbf811d3c51..21e60b1e2ff4 100644
--- a/drivers/infiniband/core/umem.c
+++ b/drivers/infiniband/core/umem.c
@@ -58,7 +58,7 @@ static void __ib_umem_release(struct ib_device *dev, struct ib_umem *umem, int d
 	for_each_sg(umem->sg_head.sgl, sg, umem->npages, i) {
 
 		page = sg_page(sg);
-		if (umem->writable && dirty)
+		if (!PageDirty(page) && umem->writable && dirty)
 			set_page_dirty_lock(page);
 		put_page(page);
 	}
diff --git a/drivers/infiniband/core/umem_odp.c b/drivers/infiniband/core/umem_odp.c
index 0780b1afefa9..8c4ec564e495 100644
--- a/drivers/infiniband/core/umem_odp.c
+++ b/drivers/infiniband/core/umem_odp.c
@@ -321,11 +321,15 @@ int ib_umem_odp_get(struct ib_ucontext *context, struct ib_umem *umem,
 		struct vm_area_struct *vma;
 		struct hstate *h;
 
+		down_read(&mm->mmap_sem);
 		vma = find_vma(mm, ib_umem_start(umem));
-		if (!vma || !is_vm_hugetlb_page(vma))
+		if (!vma || !is_vm_hugetlb_page(vma)) {
+			up_read(&mm->mmap_sem);
 			return -EINVAL;
+		}
 		h = hstate_vma(vma);
 		umem->page_shift = huge_page_shift(h);
+		up_read(&mm->mmap_sem);
 		umem->hugetlb = 1;
 	} else {
 		umem->hugetlb = 0;
diff --git a/drivers/infiniband/core/uverbs_marshall.c b/drivers/infiniband/core/uverbs_marshall.c
index 8b9587fe2303..94fd989c9060 100644
--- a/drivers/infiniband/core/uverbs_marshall.c
+++ b/drivers/infiniband/core/uverbs_marshall.c
@@ -96,11 +96,11 @@ void ib_copy_qp_attr_to_user(struct ib_uverbs_qp_attr *dst,
 }
 EXPORT_SYMBOL(ib_copy_qp_attr_to_user);
 
-void __ib_copy_path_rec_to_user(struct ib_user_path_rec *dst,
-				struct sa_path_rec *src)
+static void __ib_copy_path_rec_to_user(struct ib_user_path_rec *dst,
+				       struct sa_path_rec *src)
 {
-	memcpy(dst->dgid, src->dgid.raw, sizeof src->dgid);
-	memcpy(dst->sgid, src->sgid.raw, sizeof src->sgid);
+	memcpy(dst->dgid, src->dgid.raw, sizeof(src->dgid));
+	memcpy(dst->sgid, src->sgid.raw, sizeof(src->sgid));
 
 	dst->dlid		= htons(ntohl(sa_path_get_dlid(src)));
 	dst->slid		= htons(ntohl(sa_path_get_slid(src)));
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index b6fe45924c6e..0910faf3587b 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -488,6 +488,7 @@ static int _put_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb)
 
 	ep = *((struct c4iw_ep **)(skb->cb + 2 * sizeof(void *)));
 	release_ep_resources(ep);
+	kfree_skb(skb);
 	return 0;
 }
 
@@ -498,6 +499,7 @@ static int _put_pass_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb)
 	ep = *((struct c4iw_ep **)(skb->cb + 2 * sizeof(void *)));
 	c4iw_put_ep(&ep->parent_ep->com);
 	release_ep_resources(ep);
+	kfree_skb(skb);
 	return 0;
 }
 
@@ -569,11 +571,13 @@ static void abort_arp_failure(void *handle, struct sk_buff *skb)
 
 	pr_debug("%s rdev %p\n", __func__, rdev);
 	req->cmd = CPL_ABORT_NO_RST;
+	skb_get(skb);
 	ret = c4iw_ofld_send(rdev, skb);
 	if (ret) {
 		__state_set(&ep->com, DEAD);
 		queue_arp_failure_cpl(ep, skb, FAKE_CPL_PUT_EP_SAFE);
-	}
+	} else
+		kfree_skb(skb);
 }
 
 static int send_flowc(struct c4iw_ep *ep)
@@ -2517,7 +2521,8 @@ static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb)
 		goto reject;
 	}
 
-	hdrs = sizeof(struct iphdr) + sizeof(struct tcphdr) +
+	hdrs = ((iptype == 4) ? sizeof(struct iphdr) : sizeof(struct ipv6hdr)) +
+	       sizeof(struct tcphdr) +
 	       ((enable_tcp_timestamps && req->tcpopt.tstamp) ? 12 : 0);
 	if (peer_mss && child_ep->mtu > (peer_mss + hdrs))
 		child_ep->mtu = peer_mss + hdrs;
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index 329fb65e8fb0..f96a96dbcf1f 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -971,7 +971,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 		 devp->rdev.lldi.sge_egrstatuspagesize);
 
 	devp->rdev.hw_queue.t4_eq_status_entries =
-		devp->rdev.lldi.sge_ingpadboundary > 64 ? 2 : 1;
+		devp->rdev.lldi.sge_egrstatuspagesize / 64;
 	devp->rdev.hw_queue.t4_max_eq_size = 65520;
 	devp->rdev.hw_queue.t4_max_iq_size = 65520;
 	devp->rdev.hw_queue.t4_max_rq_size = 8192 -
diff --git a/drivers/infiniband/hw/hfi1/chip.c b/drivers/infiniband/hw/hfi1/chip.c
index 5d6b1eeaa9a0..2ba00b89df6a 100644
--- a/drivers/infiniband/hw/hfi1/chip.c
+++ b/drivers/infiniband/hw/hfi1/chip.c
@@ -6312,25 +6312,38 @@ static void handle_8051_request(struct hfi1_pportdata *ppd)
 	}
 }
 
-static void write_global_credit(struct hfi1_devdata *dd,
-				u8 vau, u16 total, u16 shared)
+/*
+ * Set up allocation unit vaulue.
+ */
+void set_up_vau(struct hfi1_devdata *dd, u8 vau)
 {
-	write_csr(dd, SEND_CM_GLOBAL_CREDIT,
-		  ((u64)total <<
-		   SEND_CM_GLOBAL_CREDIT_TOTAL_CREDIT_LIMIT_SHIFT) |
-		  ((u64)shared <<
-		   SEND_CM_GLOBAL_CREDIT_SHARED_LIMIT_SHIFT) |
-		  ((u64)vau << SEND_CM_GLOBAL_CREDIT_AU_SHIFT));
+	u64 reg = read_csr(dd, SEND_CM_GLOBAL_CREDIT);
+
+	/* do not modify other values in the register */
+	reg &= ~SEND_CM_GLOBAL_CREDIT_AU_SMASK;
+	reg |= (u64)vau << SEND_CM_GLOBAL_CREDIT_AU_SHIFT;
+	write_csr(dd, SEND_CM_GLOBAL_CREDIT, reg);
 }
 
 /*
  * Set up initial VL15 credits of the remote.  Assumes the rest of
- * the CM credit registers are zero from a previous global or credit reset .
+ * the CM credit registers are zero from a previous global or credit reset.
+ * Shared limit for VL15 will always be 0.
  */
-void set_up_vl15(struct hfi1_devdata *dd, u8 vau, u16 vl15buf)
+void set_up_vl15(struct hfi1_devdata *dd, u16 vl15buf)
 {
-	/* leave shared count at zero for both global and VL15 */
-	write_global_credit(dd, vau, vl15buf, 0);
+	u64 reg = read_csr(dd, SEND_CM_GLOBAL_CREDIT);
+
+	/* set initial values for total and shared credit limit */
+	reg &= ~(SEND_CM_GLOBAL_CREDIT_TOTAL_CREDIT_LIMIT_SMASK |
+		 SEND_CM_GLOBAL_CREDIT_SHARED_LIMIT_SMASK);
+
+	/*
+	 * Set total limit to be equal to VL15 credits.
+	 * Leave shared limit at 0.
+	 */
+	reg |= (u64)vl15buf << SEND_CM_GLOBAL_CREDIT_TOTAL_CREDIT_LIMIT_SHIFT;
+	write_csr(dd, SEND_CM_GLOBAL_CREDIT, reg);
 
 	write_csr(dd, SEND_CM_CREDIT_VL15, (u64)vl15buf
 		  << SEND_CM_CREDIT_VL15_DEDICATED_LIMIT_VL_SHIFT);
@@ -6348,9 +6361,11 @@ void reset_link_credits(struct hfi1_devdata *dd)
 	for (i = 0; i < TXE_NUM_DATA_VL; i++)
 		write_csr(dd, SEND_CM_CREDIT_VL + (8 * i), 0);
 	write_csr(dd, SEND_CM_CREDIT_VL15, 0);
-	write_global_credit(dd, 0, 0, 0);
+	write_csr(dd, SEND_CM_GLOBAL_CREDIT, 0);
 	/* reset the CM block */
 	pio_send_control(dd, PSC_CM_RESET);
+	/* reset cached value */
+	dd->vl15buf_cached = 0;
 }
 
 /* convert a vCU to a CU */
@@ -6839,24 +6854,35 @@ void handle_link_up(struct work_struct *work)
 {
 	struct hfi1_pportdata *ppd = container_of(work, struct hfi1_pportdata,
 						  link_up_work);
+	struct hfi1_devdata *dd = ppd->dd;
+
 	set_link_state(ppd, HLS_UP_INIT);
 
 	/* cache the read of DC_LCB_STS_ROUND_TRIP_LTP_CNT */
-	read_ltp_rtt(ppd->dd);
+	read_ltp_rtt(dd);
 	/*
 	 * OPA specifies that certain counters are cleared on a transition
 	 * to link up, so do that.
 	 */
-	clear_linkup_counters(ppd->dd);
+	clear_linkup_counters(dd);
 	/*
 	 * And (re)set link up default values.
 	 */
 	set_linkup_defaults(ppd);
 
+	/*
+	 * Set VL15 credits. Use cached value from verify cap interrupt.
+	 * In case of quick linkup or simulator, vl15 value will be set by
+	 * handle_linkup_change. VerifyCap interrupt handler will not be
+	 * called in those scenarios.
+	 */
+	if (!(quick_linkup || dd->icode == ICODE_FUNCTIONAL_SIMULATOR))
+		set_up_vl15(dd, dd->vl15buf_cached);
+
 	/* enforce link speed enabled */
 	if ((ppd->link_speed_active & ppd->link_speed_enabled) == 0) {
 		/* oops - current speed is not enabled, bounce */
-		dd_dev_err(ppd->dd,
+		dd_dev_err(dd,
 			   "Link speed active 0x%x is outside enabled 0x%x, downing link\n",
 			   ppd->link_speed_active, ppd->link_speed_enabled);
 		set_link_down_reason(ppd, OPA_LINKDOWN_REASON_SPEED_POLICY, 0,
@@ -7357,7 +7383,14 @@ void handle_verify_cap(struct work_struct *work)
 	 */
 	if (vau == 0)
 		vau = 1;
-	set_up_vl15(dd, vau, vl15buf);
+	set_up_vau(dd, vau);
+
+	/*
+	 * Set VL15 credits to 0 in global credit register. Cache remote VL15
+	 * credits value and wait for link-up interrupt ot set it.
+	 */
+	set_up_vl15(dd, 0);
+	dd->vl15buf_cached = vl15buf;
 
 	/* set up the LCB CRC mode */
 	crc_mask = ppd->port_crc_mode_enabled & partner_supported_crc;
diff --git a/drivers/infiniband/hw/hfi1/chip_registers.h b/drivers/infiniband/hw/hfi1/chip_registers.h
index 5bfa839d1c48..793514f1d15f 100644
--- a/drivers/infiniband/hw/hfi1/chip_registers.h
+++ b/drivers/infiniband/hw/hfi1/chip_registers.h
@@ -839,7 +839,9 @@
 #define SEND_CM_CTRL_FORCE_CREDIT_MODE_SMASK 0x8ull
 #define SEND_CM_CTRL_RESETCSR 0x0000000000000020ull
 #define SEND_CM_GLOBAL_CREDIT (TXE + 0x000000000508)
+#define SEND_CM_GLOBAL_CREDIT_AU_MASK 0x7ull
 #define SEND_CM_GLOBAL_CREDIT_AU_SHIFT 16
+#define SEND_CM_GLOBAL_CREDIT_AU_SMASK 0x70000ull
 #define SEND_CM_GLOBAL_CREDIT_RESETCSR 0x0000094000030000ull
 #define SEND_CM_GLOBAL_CREDIT_SHARED_LIMIT_MASK 0xFFFFull
 #define SEND_CM_GLOBAL_CREDIT_SHARED_LIMIT_SHIFT 0
diff --git a/drivers/infiniband/hw/hfi1/hfi.h b/drivers/infiniband/hw/hfi1/hfi.h
index da322e6668cc..414a04a481c2 100644
--- a/drivers/infiniband/hw/hfi1/hfi.h
+++ b/drivers/infiniband/hw/hfi1/hfi.h
@@ -1045,6 +1045,14 @@ struct hfi1_devdata {
 	/* initial vl15 credits to use */
 	u16 vl15_init;
 
+	/*
+	 * Cached value for vl15buf, read during verify cap interrupt. VL15
+	 * credits are to be kept at 0 and set when handling the link-up
+	 * interrupt. This removes the possibility of receiving VL15 MAD
+	 * packets before this HFI is ready.
+	 */
+	u16 vl15buf_cached;
+
 	/* Misc small ints */
 	u8 n_krcv_queues;
 	u8 qos_shift;
@@ -1598,7 +1606,8 @@ int hfi1_rcvbuf_validate(u32 size, u8 type, u16 *encode);
 int fm_get_table(struct hfi1_pportdata *ppd, int which, void *t);
 int fm_set_table(struct hfi1_pportdata *ppd, int which, void *t);
 
-void set_up_vl15(struct hfi1_devdata *dd, u8 vau, u16 vl15buf);
+void set_up_vau(struct hfi1_devdata *dd, u8 vau);
+void set_up_vl15(struct hfi1_devdata *dd, u16 vl15buf);
 void reset_link_credits(struct hfi1_devdata *dd);
 void assign_remote_cm_au_table(struct hfi1_devdata *dd, u8 vcu);
 
diff --git a/drivers/infiniband/hw/hfi1/intr.c b/drivers/infiniband/hw/hfi1/intr.c
index ba265d0ae93b..04a5082d5ac5 100644
--- a/drivers/infiniband/hw/hfi1/intr.c
+++ b/drivers/infiniband/hw/hfi1/intr.c
@@ -130,7 +130,8 @@ void handle_linkup_change(struct hfi1_devdata *dd, u32 linkup)
 		 * the remote values.  Both sides must be using the values.
 		 */
 		if (quick_linkup || dd->icode == ICODE_FUNCTIONAL_SIMULATOR) {
-			set_up_vl15(dd, dd->vau, dd->vl15_init);
+			set_up_vau(dd, dd->vau);
+			set_up_vl15(dd, dd->vl15_init);
 			assign_remote_cm_au_table(dd, dd->vcu);
 		}
 
diff --git a/drivers/infiniband/hw/hfi1/pcie.c b/drivers/infiniband/hw/hfi1/pcie.c
index 93faf86d54b6..6a9f6f9819e1 100644
--- a/drivers/infiniband/hw/hfi1/pcie.c
+++ b/drivers/infiniband/hw/hfi1/pcie.c
@@ -207,8 +207,8 @@ int hfi1_pcie_ddinit(struct hfi1_devdata *dd, struct pci_dev *pdev)
 	/*
 	 * Save BARs and command to rewrite after device reset.
 	 */
-	dd->pcibar0 = addr;
-	dd->pcibar1 = addr >> 32;
+	pci_read_config_dword(dd->pcidev, PCI_BASE_ADDRESS_0, &dd->pcibar0);
+	pci_read_config_dword(dd->pcidev, PCI_BASE_ADDRESS_1, &dd->pcibar1);
 	pci_read_config_dword(dd->pcidev, PCI_ROM_ADDRESS, &dd->pci_rom);
 	pci_read_config_word(dd->pcidev, PCI_COMMAND, &dd->pci_command);
 	pcie_capability_read_word(dd->pcidev, PCI_EXP_DEVCTL, &dd->pcie_devctl);
diff --git a/drivers/infiniband/hw/hfi1/rc.c b/drivers/infiniband/hw/hfi1/rc.c
index 069bdaf061ab..1080778a1f7c 100644
--- a/drivers/infiniband/hw/hfi1/rc.c
+++ b/drivers/infiniband/hw/hfi1/rc.c
@@ -2159,8 +2159,11 @@ send_last:
 		ret = hfi1_rvt_get_rwqe(qp, 1);
 		if (ret < 0)
 			goto nack_op_err;
-		if (!ret)
+		if (!ret) {
+			/* peer will send again */
+			rvt_put_ss(&qp->r_sge);
 			goto rnr_nak;
+		}
 		wc.ex.imm_data = ohdr->u.rc.imm_data;
 		wc.wc_flags = IB_WC_WITH_IMM;
 		goto send_last;
diff --git a/drivers/infiniband/hw/hfi1/sysfs.c b/drivers/infiniband/hw/hfi1/sysfs.c
index 50d140d25e38..2f3bbcac1e34 100644
--- a/drivers/infiniband/hw/hfi1/sysfs.c
+++ b/drivers/infiniband/hw/hfi1/sysfs.c
@@ -196,7 +196,8 @@ static const struct sysfs_ops port_cc_sysfs_ops = {
 };
 
 static struct attribute *port_cc_default_attributes[] = {
-	&cc_prescan_attr.attr
+	&cc_prescan_attr.attr,
+	NULL
 };
 
 static struct kobj_type port_cc_ktype = {
diff --git a/drivers/infiniband/hw/i40iw/i40iw_cm.c b/drivers/infiniband/hw/i40iw/i40iw_cm.c
index f3bc01bce483..6ae98aa7f74e 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_cm.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_cm.c
@@ -784,7 +784,6 @@ static void i40iw_build_mpa_v2(struct i40iw_cm_node *cm_node,
 	}
 
 	ctrl_ird |= IETF_PEER_TO_PEER;
-	ctrl_ird |= IETF_FLPDU_ZERO_LEN;
 
 	switch (mpa_key) {
 	case MPA_KEY_REQUEST:
@@ -2446,8 +2445,8 @@ static void i40iw_handle_rcv_mpa(struct i40iw_cm_node *cm_node,
 		} else {
 			type = I40IW_CM_EVENT_CONNECTED;
 			cm_node->state = I40IW_CM_STATE_OFFLOADED;
-			i40iw_send_ack(cm_node);
 		}
+		i40iw_send_ack(cm_node);
 		break;
 	default:
 		pr_err("%s wrong cm_node state =%d\n", __func__, cm_node->state);
diff --git a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
index f82483b3d1e7..a027e2072477 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_ctrl.c
@@ -285,28 +285,20 @@ void i40iw_change_l2params(struct i40iw_sc_vsi *vsi, struct i40iw_l2params *l2pa
 	struct i40iw_sc_dev *dev = vsi->dev;
 	struct i40iw_sc_qp *qp = NULL;
 	bool qs_handle_change = false;
-	bool mss_change = false;
 	unsigned long flags;
 	u16 qs_handle;
 	int i;
 
-	if (vsi->mss != l2params->mss) {
-		mss_change = true;
-		vsi->mss = l2params->mss;
-	}
+	vsi->mss = l2params->mss;
 
 	i40iw_fill_qos_list(l2params->qs_handle_list);
 	for (i = 0; i < I40IW_MAX_USER_PRIORITY; i++) {
 		qs_handle = l2params->qs_handle_list[i];
 		if (vsi->qos[i].qs_handle != qs_handle)
 			qs_handle_change = true;
-		else if (!mss_change)
-			continue;       /* no MSS nor qs handle change */
 		spin_lock_irqsave(&vsi->qos[i].lock, flags);
 		qp = i40iw_get_qp(&vsi->qos[i].qplist, qp);
 		while (qp) {
-			if (mss_change)
-				i40iw_qp_mss_modify(dev, qp);
 			if (qs_handle_change) {
 				qp->qs_handle = qs_handle;
 				/* issue cqp suspend command */
@@ -2395,7 +2387,6 @@ static enum i40iw_status_code i40iw_sc_qp_modify(
 
 	set_64bit_val(wqe,
 		      8,
-		      LS_64(info->new_mss, I40IW_CQPSQ_QP_NEWMSS) |
 		      LS_64(term_len, I40IW_CQPSQ_QP_TERMLEN));
 
 	set_64bit_val(wqe, 16, qp->hw_host_ctx_pa);
@@ -2410,7 +2401,6 @@ static enum i40iw_status_code i40iw_sc_qp_modify(
 		 LS_64(info->cq_num_valid, I40IW_CQPSQ_QP_CQNUMVALID) |
 		 LS_64(info->force_loopback, I40IW_CQPSQ_QP_FORCELOOPBACK) |
 		 LS_64(qp->qp_type, I40IW_CQPSQ_QP_QPTYPE) |
-		 LS_64(info->mss_change, I40IW_CQPSQ_QP_MSSCHANGE) |
 		 LS_64(info->static_rsrc, I40IW_CQPSQ_QP_STATRSRC) |
 		 LS_64(info->remove_hash_idx, I40IW_CQPSQ_QP_REMOVEHASHENTRY) |
 		 LS_64(term_actions, I40IW_CQPSQ_QP_TERMACT) |
diff --git a/drivers/infiniband/hw/i40iw/i40iw_main.c b/drivers/infiniband/hw/i40iw/i40iw_main.c
index 2728af3103ce..a3f18a22f5ed 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_main.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_main.c
@@ -1319,13 +1319,13 @@ static enum i40iw_status_code i40iw_initialize_dev(struct i40iw_device *iwdev,
 	status = i40iw_obj_aligned_mem(iwdev, &mem, I40IW_QUERY_FPM_BUF_SIZE,
 				       I40IW_FPM_QUERY_BUF_ALIGNMENT_MASK);
 	if (status)
-		goto exit;
+		goto error;
 	info.fpm_query_buf_pa = mem.pa;
 	info.fpm_query_buf = mem.va;
 	status = i40iw_obj_aligned_mem(iwdev, &mem, I40IW_COMMIT_FPM_BUF_SIZE,
 				       I40IW_FPM_COMMIT_BUF_ALIGNMENT_MASK);
 	if (status)
-		goto exit;
+		goto error;
 	info.fpm_commit_buf_pa = mem.pa;
 	info.fpm_commit_buf = mem.va;
 	info.hmc_fn_id = ldev->fid;
@@ -1347,11 +1347,9 @@ static enum i40iw_status_code i40iw_initialize_dev(struct i40iw_device *iwdev,
 	info.exception_lan_queue = 1;
 	info.vchnl_send = i40iw_virtchnl_send;
 	status = i40iw_device_init(&iwdev->sc_dev, &info);
-exit:
-	if (status) {
-		kfree(iwdev->hmc_info_mem);
-		iwdev->hmc_info_mem = NULL;
-	}
+
+	if (status)
+		goto error;
 	memset(&vsi_info, 0, sizeof(vsi_info));
 	vsi_info.dev = &iwdev->sc_dev;
 	vsi_info.back_vsi = (void *)iwdev;
@@ -1362,11 +1360,19 @@ exit:
 		memset(&stats_info, 0, sizeof(stats_info));
 		stats_info.fcn_id = ldev->fid;
 		stats_info.pestat = kzalloc(sizeof(*stats_info.pestat), GFP_KERNEL);
+		if (!stats_info.pestat) {
+			status = I40IW_ERR_NO_MEMORY;
+			goto error;
+		}
 		stats_info.stats_initialize = true;
 		if (stats_info.pestat)
 			i40iw_vsi_stats_init(&iwdev->vsi, &stats_info);
 	}
 	return status;
+error:
+	kfree(iwdev->hmc_info_mem);
+	iwdev->hmc_info_mem = NULL;
+	return status;
 }
 
 /**
diff --git a/drivers/infiniband/hw/i40iw/i40iw_osdep.h b/drivers/infiniband/hw/i40iw/i40iw_osdep.h
index aa66c1c63dfa..f27be3e7830b 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_osdep.h
+++ b/drivers/infiniband/hw/i40iw/i40iw_osdep.h
@@ -199,7 +199,6 @@ void i40iw_cqp_spawn_worker(struct i40iw_sc_dev *dev,
 			    struct i40iw_virtchnl_work_info *work_info, u32 iw_vf_idx);
 void *i40iw_remove_head(struct list_head *list);
 void i40iw_qp_suspend_resume(struct i40iw_sc_dev *dev, struct i40iw_sc_qp *qp, bool suspend);
-void i40iw_qp_mss_modify(struct i40iw_sc_dev *dev, struct i40iw_sc_qp *qp);
 
 void i40iw_term_modify_qp(struct i40iw_sc_qp *qp, u8 next_state, u8 term, u8 term_len);
 void i40iw_terminate_done(struct i40iw_sc_qp *qp, int timeout_occurred);
diff --git a/drivers/infiniband/hw/i40iw/i40iw_type.h b/drivers/infiniband/hw/i40iw/i40iw_type.h
index 7b76259752b0..959ec81fba99 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_type.h
+++ b/drivers/infiniband/hw/i40iw/i40iw_type.h
@@ -541,7 +541,6 @@ struct i40iw_create_qp_info {
 struct i40iw_modify_qp_info {
 	u64 rx_win0;
 	u64 rx_win1;
-	u16 new_mss;
 	u8 next_iwarp_state;
 	u8 termlen;
 	bool ord_valid;
@@ -554,7 +553,6 @@ struct i40iw_modify_qp_info {
 	bool dont_send_term;
 	bool dont_send_fin;
 	bool cached_var_valid;
-	bool mss_change;
 	bool force_loopback;
 };
 
diff --git a/drivers/infiniband/hw/i40iw/i40iw_utils.c b/drivers/infiniband/hw/i40iw/i40iw_utils.c
index 409a3781e735..56d986924a4c 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_utils.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_utils.c
@@ -757,23 +757,6 @@ void i40iw_qp_suspend_resume(struct i40iw_sc_dev *dev, struct i40iw_sc_qp *qp, b
 }
 
 /**
- * i40iw_qp_mss_modify - modify mss for qp
- * @dev: hardware control device structure
- * @qp: hardware control qp
- */
-void i40iw_qp_mss_modify(struct i40iw_sc_dev *dev, struct i40iw_sc_qp *qp)
-{
-	struct i40iw_device *iwdev = (struct i40iw_device *)dev->back_dev;
-	struct i40iw_qp *iwqp = (struct i40iw_qp *)qp->back_qp;
-	struct i40iw_modify_qp_info info;
-
-	memset(&info, 0, sizeof(info));
-	info.mss_change = true;
-	info.new_mss = qp->vsi->mss;
-	i40iw_hw_modify_qp(iwdev, iwqp, &info, false);
-}
-
-/**
  * i40iw_term_modify_qp - modify qp for term message
  * @qp: hardware control qp
  * @next_state: qp's next state
diff --git a/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c b/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
index f4d13683a403..48fd327f876b 100644
--- a/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
+++ b/drivers/infiniband/hw/i40iw/i40iw_virtchnl.c
@@ -443,10 +443,7 @@ enum i40iw_status_code i40iw_vchnl_recv_pf(struct i40iw_sc_dev *dev,
 	if (!dev->vchnl_up)
 		return I40IW_ERR_NOT_READY;
 	if (vchnl_msg->iw_op_code == I40IW_VCHNL_OP_GET_VER) {
-		if (vchnl_msg->iw_op_ver != I40IW_VCHNL_OP_GET_VER_V0)
-			vchnl_pf_send_get_ver_resp(dev, vf_id, vchnl_msg);
-		else
-			vchnl_pf_send_get_ver_resp(dev, vf_id, vchnl_msg);
+		vchnl_pf_send_get_ver_resp(dev, vf_id, vchnl_msg);
 		return I40IW_SUCCESS;
 	}
 	for (iw_vf_idx = 0; iw_vf_idx < I40IW_MAX_PE_ENABLED_VF_COUNT; iw_vf_idx++) {
diff --git a/drivers/infiniband/hw/mlx4/mad.c b/drivers/infiniband/hw/mlx4/mad.c
index b4694717f6f3..21d31cb1325f 100644
--- a/drivers/infiniband/hw/mlx4/mad.c
+++ b/drivers/infiniband/hw/mlx4/mad.c
@@ -1578,6 +1578,7 @@ static void mlx4_ib_multiplex_mad(struct mlx4_ib_demux_pv_ctx *ctx, struct ib_wc
 	if (port < 0)
 		return;
 	ah.av.ib.port_pd = cpu_to_be32(port << 24 | (be32_to_cpu(ah.av.ib.port_pd) & 0xffffff));
+	ah.ibah.type = rdma_ah_find_type(&dev->ib_dev, port);
 
 	mlx4_ib_query_ah(&ah.ibah, &ah_attr);
 	if (rdma_ah_get_ah_flags(&ah_attr) & IB_AH_GRH)
diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index d45772da0963..0c79983c8b1a 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -2979,6 +2979,18 @@ error_0:
 	return ret;
 }
 
+static u8 mlx5_get_umr_fence(u8 umr_fence_cap)
+{
+	switch (umr_fence_cap) {
+	case MLX5_CAP_UMR_FENCE_NONE:
+		return MLX5_FENCE_MODE_NONE;
+	case MLX5_CAP_UMR_FENCE_SMALL:
+		return MLX5_FENCE_MODE_INITIATOR_SMALL;
+	default:
+		return MLX5_FENCE_MODE_STRONG_ORDERING;
+	}
+}
+
 static int create_dev_resources(struct mlx5_ib_resources *devr)
 {
 	struct ib_srq_init_attr attr;
@@ -3693,6 +3705,8 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 
 	mlx5_ib_internal_fill_odp_caps(dev);
 
+	dev->umr_fence = mlx5_get_umr_fence(MLX5_CAP_GEN(mdev, umr_fence));
+
 	if (MLX5_CAP_GEN(mdev, imaicl)) {
 		dev->ib_dev.alloc_mw		= mlx5_ib_alloc_mw;
 		dev->ib_dev.dealloc_mw		= mlx5_ib_dealloc_mw;
diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
index 38c877bc45e5..bdcf25410c99 100644
--- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
+++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
@@ -349,7 +349,7 @@ struct mlx5_ib_qp {
 	struct mlx5_ib_wq	rq;
 
 	u8			sq_signal_bits;
-	u8			fm_cache;
+	u8			next_fence;
 	struct mlx5_ib_wq	sq;
 
 	/* serialize qp state modifications
@@ -654,6 +654,7 @@ struct mlx5_ib_dev {
 	struct mlx5_ib_port	*port;
 	struct mlx5_sq_bfreg     bfreg;
 	struct mlx5_sq_bfreg     fp_bfreg;
+	u8				umr_fence;
 };
 
 static inline struct mlx5_ib_cq *to_mibcq(struct mlx5_core_cq *mcq)
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index 93959e1e43a3..ebb6768684de 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -3738,24 +3738,6 @@ static void dump_wqe(struct mlx5_ib_qp *qp, int idx, int size_16)
 	}
 }
 
-static u8 get_fence(u8 fence, struct ib_send_wr *wr)
-{
-	if (unlikely(wr->opcode == IB_WR_LOCAL_INV &&
-		     wr->send_flags & IB_SEND_FENCE))
-		return MLX5_FENCE_MODE_STRONG_ORDERING;
-
-	if (unlikely(fence)) {
-		if (wr->send_flags & IB_SEND_FENCE)
-			return MLX5_FENCE_MODE_SMALL_AND_FENCE;
-		else
-			return fence;
-	} else if (unlikely(wr->send_flags & IB_SEND_FENCE)) {
-		return MLX5_FENCE_MODE_FENCE;
-	}
-
-	return 0;
-}
-
 static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
 		     struct mlx5_wqe_ctrl_seg **ctrl,
 		     struct ib_send_wr *wr, unsigned *idx,
@@ -3784,8 +3766,7 @@ static int begin_wqe(struct mlx5_ib_qp *qp, void **seg,
 static void finish_wqe(struct mlx5_ib_qp *qp,
 		       struct mlx5_wqe_ctrl_seg *ctrl,
 		       u8 size, unsigned idx, u64 wr_id,
-		       int nreq, u8 fence, u8 next_fence,
-		       u32 mlx5_opcode)
+		       int nreq, u8 fence, u32 mlx5_opcode)
 {
 	u8 opmod = 0;
 
@@ -3793,7 +3774,6 @@ static void finish_wqe(struct mlx5_ib_qp *qp,
 					     mlx5_opcode | ((u32)opmod << 24));
 	ctrl->qpn_ds = cpu_to_be32(size | (qp->trans_qp.base.mqp.qpn << 8));
 	ctrl->fm_ce_se |= fence;
-	qp->fm_cache = next_fence;
 	if (unlikely(qp->wq_sig))
 		ctrl->signature = wq_sig(ctrl);
 
@@ -3853,7 +3833,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			goto out;
 		}
 
-		fence = qp->fm_cache;
 		num_sge = wr->num_sge;
 		if (unlikely(num_sge > qp->sq.max_gs)) {
 			mlx5_ib_warn(dev, "\n");
@@ -3870,6 +3849,19 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			goto out;
 		}
 
+		if (wr->opcode == IB_WR_LOCAL_INV ||
+		    wr->opcode == IB_WR_REG_MR) {
+			fence = dev->umr_fence;
+			next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
+		} else if (wr->send_flags & IB_SEND_FENCE) {
+			if (qp->next_fence)
+				fence = MLX5_FENCE_MODE_SMALL_AND_FENCE;
+			else
+				fence = MLX5_FENCE_MODE_FENCE;
+		} else {
+			fence = qp->next_fence;
+		}
+
 		switch (ibqp->qp_type) {
 		case IB_QPT_XRC_INI:
 			xrc = seg;
@@ -3896,7 +3888,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				goto out;
 
 			case IB_WR_LOCAL_INV:
-				next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
 				qp->sq.wr_data[idx] = IB_WR_LOCAL_INV;
 				ctrl->imm = cpu_to_be32(wr->ex.invalidate_rkey);
 				set_linv_wr(qp, &seg, &size);
@@ -3904,7 +3895,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				break;
 
 			case IB_WR_REG_MR:
-				next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
 				qp->sq.wr_data[idx] = IB_WR_REG_MR;
 				ctrl->imm = cpu_to_be32(reg_wr(wr)->key);
 				err = set_reg_wr(qp, reg_wr(wr), &seg, &size);
@@ -3927,9 +3917,8 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 					goto out;
 				}
 
-				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
-					   nreq, get_fence(fence, wr),
-					   next_fence, MLX5_OPCODE_UMR);
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
+					   fence, MLX5_OPCODE_UMR);
 				/*
 				 * SET_PSV WQEs are not signaled and solicited
 				 * on error
@@ -3954,9 +3943,8 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 					goto out;
 				}
 
-				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
-					   nreq, get_fence(fence, wr),
-					   next_fence, MLX5_OPCODE_SET_PSV);
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
+					   fence, MLX5_OPCODE_SET_PSV);
 				err = begin_wqe(qp, &seg, &ctrl, wr,
 						&idx, &size, nreq);
 				if (err) {
@@ -3966,7 +3954,6 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 					goto out;
 				}
 
-				next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
 				err = set_psv_wr(&sig_handover_wr(wr)->sig_attrs->wire,
 						 mr->sig->psv_wire.psv_idx, &seg,
 						 &size);
@@ -3976,9 +3963,9 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 					goto out;
 				}
 
-				finish_wqe(qp, ctrl, size, idx, wr->wr_id,
-					   nreq, get_fence(fence, wr),
-					   next_fence, MLX5_OPCODE_SET_PSV);
+				finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
+					   fence, MLX5_OPCODE_SET_PSV);
+				qp->next_fence = MLX5_FENCE_MODE_INITIATOR_SMALL;
 				num_sge = 0;
 				goto skip_psv;
 
@@ -4089,8 +4076,8 @@ int mlx5_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			}
 		}
 
-		finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq,
-			   get_fence(fence, wr), next_fence,
+		qp->next_fence = next_fence;
+		finish_wqe(qp, ctrl, size, idx, wr->wr_id, nreq, fence,
 			   mlx5_ib_opcode[wr->opcode]);
 skip_psv:
 		if (0)
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c
index fb983df7c157..30b256a2c54e 100644
--- a/drivers/infiniband/hw/nes/nes_cm.c
+++ b/drivers/infiniband/hw/nes/nes_cm.c
@@ -610,7 +610,6 @@ static void build_mpa_v2(struct nes_cm_node *cm_node,
 		ctrl_ord = cm_node->ord_size & IETF_NO_IRD_ORD;
 	}
 	ctrl_ird |= IETF_PEER_TO_PEER;
-	ctrl_ird |= IETF_FLPDU_ZERO_LEN;
 
 	switch (mpa_key) {
 	case MPA_KEY_REQUEST:
@@ -1826,7 +1825,7 @@ static void handle_rcv_mpa(struct nes_cm_node *cm_node, struct sk_buff *skb)
 			type = NES_CM_EVENT_CONNECTED;
 			cm_node->state = NES_CM_STATE_TSA;
 		}
-
+		send_ack(cm_node, NULL);
 		break;
 	default:
 		WARN_ON(1);
diff --git a/drivers/infiniband/hw/qedr/qedr_cm.c b/drivers/infiniband/hw/qedr/qedr_cm.c
index 3d7705cec770..d86dbe814d98 100644
--- a/drivers/infiniband/hw/qedr/qedr_cm.c
+++ b/drivers/infiniband/hw/qedr/qedr_cm.c
@@ -270,11 +270,13 @@ static inline int qedr_gsi_build_header(struct qedr_dev *dev,
 		return rc;
 	}
 
-	vlan_id = rdma_vlan_dev_vlan_id(sgid_attr.ndev);
-	if (vlan_id < VLAN_CFI_MASK)
-		has_vlan = true;
-	if (sgid_attr.ndev)
+	if (sgid_attr.ndev) {
+		vlan_id = rdma_vlan_dev_vlan_id(sgid_attr.ndev);
+		if (vlan_id < VLAN_CFI_MASK)
+			has_vlan = true;
+
 		dev_put(sgid_attr.ndev);
+	}
 
 	if (!memcmp(&sgid, &zgid, sizeof(sgid))) {
 		DP_ERR(dev, "gsi post send: GID not found GID index %d\n",
diff --git a/drivers/infiniband/hw/qib/qib_rc.c b/drivers/infiniband/hw/qib/qib_rc.c
index fc8b88514da5..4ddbcac5eabe 100644
--- a/drivers/infiniband/hw/qib/qib_rc.c
+++ b/drivers/infiniband/hw/qib/qib_rc.c
@@ -1956,8 +1956,10 @@ send_last:
 		ret = qib_get_rwqe(qp, 1);
 		if (ret < 0)
 			goto nack_op_err;
-		if (!ret)
+		if (!ret) {
+			rvt_put_ss(&qp->r_sge);
 			goto rnr_nak;
+		}
 		wc.ex.imm_data = ohdr->u.rc.imm_data;
 		hdrsize += 4;
 		wc.wc_flags = IB_WC_WITH_IMM;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
index 874b24366e4d..7871379342f4 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_ethtool.c
@@ -178,7 +178,7 @@ static inline int ib_speed_enum_to_int(int speed)
 static int ipoib_get_link_ksettings(struct net_device *netdev,
 				    struct ethtool_link_ksettings *cmd)
 {
-	struct ipoib_dev_priv *priv = netdev_priv(netdev);
+	struct ipoib_dev_priv *priv = ipoib_priv(netdev);
 	struct ib_port_attr attr;
 	int ret, speed, width;
 
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 2869d1adb1de..a115c0b7a310 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -1590,7 +1590,7 @@ static void ipoib_neigh_hash_uninit(struct net_device *dev)
 	wait_for_completion(&priv->ntbl.deleted);
 }
 
-void ipoib_dev_uninit_default(struct net_device *dev)
+static void ipoib_dev_uninit_default(struct net_device *dev)
 {
 	struct ipoib_dev_priv *priv = ipoib_priv(dev);
 
diff --git a/drivers/infiniband/ulp/srp/ib_srp.c b/drivers/infiniband/ulp/srp/ib_srp.c
index def723a5df29..2354c742caa1 100644
--- a/drivers/infiniband/ulp/srp/ib_srp.c
+++ b/drivers/infiniband/ulp/srp/ib_srp.c
@@ -320,7 +320,7 @@ static int srp_new_cm_id(struct srp_rdma_ch *ch)
 	ch->path.sgid = target->sgid;
 	ch->path.dgid = target->orig_dgid;
 	ch->path.pkey = target->pkey;
-	sa_path_set_service_id(&ch->path, target->service_id);
+	ch->path.service_id = target->service_id;
 
 	return 0;
 }
@@ -575,7 +575,7 @@ static int srp_create_ch_ib(struct srp_rdma_ch *ch)
 	return 0;
 
 err_qp:
-	srp_destroy_qp(ch, qp);
+	ib_destroy_qp(qp);
 
 err_send_cq:
 	ib_free_cq(send_cq);
diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h
index 32de0724b400..edafedb7b509 100644
--- a/include/linux/mlx5/mlx5_ifc.h
+++ b/include/linux/mlx5/mlx5_ifc.h
@@ -766,6 +766,12 @@ enum {
 	MLX5_CAP_PORT_TYPE_ETH = 0x1,
 };
 
+enum {
+	MLX5_CAP_UMR_FENCE_STRONG	= 0x0,
+	MLX5_CAP_UMR_FENCE_SMALL	= 0x1,
+	MLX5_CAP_UMR_FENCE_NONE		= 0x2,
+};
+
 struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         reserved_at_0[0x80];
 
@@ -875,7 +881,9 @@ struct mlx5_ifc_cmd_hca_cap_bits {
 	u8         reserved_at_202[0x1];
 	u8         ipoib_enhanced_offloads[0x1];
 	u8         ipoib_basic_offloads[0x1];
-	u8         reserved_at_205[0xa];
+	u8         reserved_at_205[0x5];
+	u8         umr_fence[0x2];
+	u8         reserved_at_20c[0x3];
 	u8         drain_sigerr[0x1];
 	u8         cmdif_checksum[0x2];
 	u8         sigerr_cqe[0x1];
diff --git a/include/rdma/ib_sa.h b/include/rdma/ib_sa.h
index f5f70e345318..355b81f4242d 100644
--- a/include/rdma/ib_sa.h
+++ b/include/rdma/ib_sa.h
@@ -158,7 +158,6 @@ enum sa_path_rec_type {
 };
 
 struct sa_path_rec_ib {
-	__be64       service_id;
 	__be16       dlid;
 	__be16       slid;
 	u8           raw_traffic;
@@ -174,7 +173,6 @@ struct sa_path_rec_roce {
 };
 
 struct sa_path_rec_opa {
-	__be64       service_id;
 	__be32       dlid;
 	__be32       slid;
 	u8           raw_traffic;
@@ -189,6 +187,7 @@ struct sa_path_rec_opa {
 struct sa_path_rec {
 	union ib_gid dgid;
 	union ib_gid sgid;
+	__be64       service_id;
 	/* reserved */
 	__be32       flow_label;
 	u8           hop_limit;
@@ -262,7 +261,7 @@ static inline void path_conv_opa_to_ib(struct sa_path_rec *ib,
 		ib->ib.dlid	= htons(ntohl(opa->opa.dlid));
 		ib->ib.slid	= htons(ntohl(opa->opa.slid));
 	}
-	ib->ib.service_id	= opa->opa.service_id;
+	ib->service_id		= opa->service_id;
 	ib->ib.raw_traffic	= opa->opa.raw_traffic;
 }
 
@@ -281,7 +280,7 @@ static inline void path_conv_ib_to_opa(struct sa_path_rec *opa,
 	}
 	opa->opa.slid		= slid;
 	opa->opa.dlid		= dlid;
-	opa->opa.service_id	= ib->ib.service_id;
+	opa->service_id		= ib->service_id;
 	opa->opa.raw_traffic	= ib->ib.raw_traffic;
 }
 
@@ -591,15 +590,6 @@ static inline bool sa_path_is_roce(struct sa_path_rec *rec)
 		(rec->rec_type == SA_PATH_REC_TYPE_ROCE_V2));
 }
 
-static inline void sa_path_set_service_id(struct sa_path_rec *rec,
-					  __be64 service_id)
-{
-	if (rec->rec_type == SA_PATH_REC_TYPE_IB)
-		rec->ib.service_id = service_id;
-	else if (rec->rec_type == SA_PATH_REC_TYPE_OPA)
-		rec->opa.service_id = service_id;
-}
-
 static inline void sa_path_set_slid(struct sa_path_rec *rec, __be32 slid)
 {
 	if (rec->rec_type == SA_PATH_REC_TYPE_IB)
@@ -625,15 +615,6 @@ static inline void sa_path_set_raw_traffic(struct sa_path_rec *rec,
 		rec->opa.raw_traffic = raw_traffic;
 }
 
-static inline __be64 sa_path_get_service_id(struct sa_path_rec *rec)
-{
-	if (rec->rec_type == SA_PATH_REC_TYPE_IB)
-		return rec->ib.service_id;
-	else if (rec->rec_type == SA_PATH_REC_TYPE_OPA)
-		return rec->opa.service_id;
-	return 0;
-}
-
 static inline __be32 sa_path_get_slid(struct sa_path_rec *rec)
 {
 	if (rec->rec_type == SA_PATH_REC_TYPE_IB)
diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h
index 585266144329..348c102cb5f6 100644
--- a/include/rdma/rdma_netlink.h
+++ b/include/rdma/rdma_netlink.h
@@ -10,9 +10,6 @@ struct ibnl_client_cbs {
 	struct module *module;
 };
 
-int ibnl_init(void);
-void ibnl_cleanup(void);
-
 /**
  * Add a a client to the list of IB netlink exporters.
  * @index: Index of the added client
@@ -77,11 +74,4 @@ int ibnl_unicast(struct sk_buff *skb, struct nlmsghdr *nlh,
 int ibnl_multicast(struct sk_buff *skb, struct nlmsghdr *nlh,
 			unsigned int group, gfp_t flags);
 
-/**
- * Check if there are any listeners to the netlink group
- * @group: the netlink group ID
- * Returns 0 on success or a negative for no listeners.
- */
-int ibnl_chk_listeners(unsigned int group);
-
 #endif /* _RDMA_NETLINK_H */