Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR (net-6.14-rc5).

Conflicts:

drivers/net/ethernet/cadence/macb_main.c
  fa52f15c745c ("net: cadence: macb: Synchronize stats calculations")
  75696dd0fd72 ("net: cadence: macb: Convert to get_stats64")
https://lore.kernel.org/20250224125848.68ee63e5@canb.auug.org.au

Adjacent changes:

drivers/net/ethernet/intel/ice/ice_sriov.c
  79990cf5e7ad ("ice: Fix deinitializing VF in error path")
  a203163274a4 ("ice: simplify VF MSI-X managing")

net/ipv4/tcp.c
  18912c520674 ("tcp: devmem: don't write truncated dmabuf CMSGs to userspace")
  297d389e9e5b ("net: prefix devmem specific helpers")

net/mptcp/subflow.c
  8668860b0ad3 ("mptcp: reset when MPTCP opts are dropped after join")
  c3349a22c200 ("mptcp: consolidate subflow cleanup")

Signed-off-by: Jakub Kicinski <kuba@kernel.org>

15 files changed, 160 insertions, 67 deletions

diff --git a/drivers/infiniband/hw/bnxt_re/bnxt_re.h b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
index b91a85a491d0..3721446c6ba4 100644
--- a/drivers/infiniband/hw/bnxt_re/bnxt_re.h
+++ b/drivers/infiniband/hw/bnxt_re/bnxt_re.h
@@ -187,7 +187,6 @@ struct bnxt_re_dev {
 #define BNXT_RE_FLAG_ISSUE_ROCE_STATS          29
 	struct net_device		*netdev;
 	struct auxiliary_device         *adev;
-	struct notifier_block		nb;
 	unsigned int			version, major, minor;
 	struct bnxt_qplib_chip_ctx	*chip_ctx;
 	struct bnxt_en_dev		*en_dev;
diff --git a/drivers/infiniband/hw/bnxt_re/hw_counters.c b/drivers/infiniband/hw/bnxt_re/hw_counters.c
index 3ac47f4e6122..f039aefcaf67 100644
--- a/drivers/infiniband/hw/bnxt_re/hw_counters.c
+++ b/drivers/infiniband/hw/bnxt_re/hw_counters.c
@@ -348,8 +348,8 @@ int bnxt_re_ib_get_hw_stats(struct ib_device *ibdev,
 			goto done;
 		}
 		bnxt_re_copy_err_stats(rdev, stats, err_s);
-		if (_is_ext_stats_supported(rdev->dev_attr->dev_cap_flags) &&
-		    !rdev->is_virtfn) {
+		if (bnxt_ext_stats_supported(rdev->chip_ctx, rdev->dev_attr->dev_cap_flags,
+					     rdev->is_virtfn)) {
 			rc = bnxt_re_get_ext_stat(rdev, stats);
 			if (rc) {
 				clear_bit(BNXT_RE_FLAG_ISSUE_ROCE_STATS,
diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
index 2de101d6e825..6f5db32082dd 100644
--- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c
+++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
@@ -1870,6 +1870,8 @@ int bnxt_re_create_srq(struct ib_srq *ib_srq,
 	srq->qplib_srq.threshold = srq_init_attr->attr.srq_limit;
 	srq->srq_limit = srq_init_attr->attr.srq_limit;
 	srq->qplib_srq.eventq_hw_ring_id = rdev->nqr->nq[0].ring_id;
+	srq->qplib_srq.sg_info.pgsize = PAGE_SIZE;
+	srq->qplib_srq.sg_info.pgshft = PAGE_SHIFT;
 	nq = &rdev->nqr->nq[0];
 
 	if (udata) {
diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c
index e9e4da4dd576..a94c8c5387d9 100644
--- a/drivers/infiniband/hw/bnxt_re/main.c
+++ b/drivers/infiniband/hw/bnxt_re/main.c
@@ -396,11 +396,16 @@ free_dcb:
 
 static void bnxt_re_async_notifier(void *handle, struct hwrm_async_event_cmpl *cmpl)
 {
-	struct bnxt_re_dev *rdev = (struct bnxt_re_dev *)handle;
+	struct bnxt_re_en_dev_info *en_info = auxiliary_get_drvdata(handle);
 	struct bnxt_re_dcb_work *dcb_work;
+	struct bnxt_re_dev *rdev;
 	u32 data1, data2;
 	u16 event_id;
 
+	rdev = en_info->rdev;
+	if (!rdev)
+		return;
+
 	event_id = le16_to_cpu(cmpl->event_id);
 	data1 = le32_to_cpu(cmpl->event_data1);
 	data2 = le32_to_cpu(cmpl->event_data2);
@@ -433,6 +438,8 @@ static void bnxt_re_stop_irq(void *handle, bool reset)
 	int indx;
 
 	rdev = en_info->rdev;
+	if (!rdev)
+		return;
 	rcfw = &rdev->rcfw;
 
 	if (reset) {
@@ -461,6 +468,8 @@ static void bnxt_re_start_irq(void *handle, struct bnxt_msix_entry *ent)
 	int indx, rc;
 
 	rdev = en_info->rdev;
+	if (!rdev)
+		return;
 	msix_ent = rdev->nqr->msix_entries;
 	rcfw = &rdev->rcfw;
 	if (!ent) {
@@ -1350,7 +1359,6 @@ static struct bnxt_re_dev *bnxt_re_dev_add(struct auxiliary_device *adev,
 		return NULL;
 	}
 	/* Default values */
-	rdev->nb.notifier_call = NULL;
 	rdev->netdev = en_dev->net;
 	rdev->en_dev = en_dev;
 	rdev->adev = adev;
@@ -2345,15 +2353,6 @@ exit:
 static void bnxt_re_remove_device(struct bnxt_re_dev *rdev, u8 op_type,
 				  struct auxiliary_device *aux_dev)
 {
-	if (rdev->nb.notifier_call) {
-		unregister_netdevice_notifier(&rdev->nb);
-		rdev->nb.notifier_call = NULL;
-	} else {
-		/* If notifier is null, we should have already done a
-		 * clean up before coming here.
-		 */
-		return;
-	}
 	bnxt_re_setup_cc(rdev, false);
 	ib_unregister_device(&rdev->ibdev);
 	bnxt_re_dev_uninit(rdev, op_type);
@@ -2433,6 +2432,7 @@ static int bnxt_re_suspend(struct auxiliary_device *adev, pm_message_t state)
 	ibdev_info(&rdev->ibdev, "%s: L2 driver notified to stop en_state 0x%lx",
 		   __func__, en_dev->en_state);
 	bnxt_re_remove_device(rdev, BNXT_RE_PRE_RECOVERY_REMOVE, adev);
+	bnxt_re_update_en_info_rdev(NULL, en_info, adev);
 	mutex_unlock(&bnxt_re_mutex);
 
 	return 0;
diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.h b/drivers/infiniband/hw/bnxt_re/qplib_res.h
index be5d907a036b..711990232de1 100644
--- a/drivers/infiniband/hw/bnxt_re/qplib_res.h
+++ b/drivers/infiniband/hw/bnxt_re/qplib_res.h
@@ -547,6 +547,14 @@ static inline bool _is_ext_stats_supported(u16 dev_cap_flags)
 		CREQ_QUERY_FUNC_RESP_SB_EXT_STATS;
 }
 
+static inline int bnxt_ext_stats_supported(struct bnxt_qplib_chip_ctx *ctx,
+					   u16 flags, bool virtfn)
+{
+	/* ext stats supported if cap flag is set AND is a PF OR a Thor2 VF */
+	return (_is_ext_stats_supported(flags) &&
+		((virtfn && bnxt_qplib_is_chip_gen_p7(ctx)) || (!virtfn)));
+}
+
 static inline bool _is_hw_retx_supported(u16 dev_cap_flags)
 {
 	return dev_cap_flags &
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
index dded339802b3..160e8927d364 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
@@ -1286,10 +1286,8 @@ static u32 hns_roce_cmdq_tx_timeout(u16 opcode, u32 tx_timeout)
 	return tx_timeout;
 }
 
-static void hns_roce_wait_csq_done(struct hns_roce_dev *hr_dev, u16 opcode)
+static void hns_roce_wait_csq_done(struct hns_roce_dev *hr_dev, u32 tx_timeout)
 {
-	struct hns_roce_v2_priv *priv = hr_dev->priv;
-	u32 tx_timeout = hns_roce_cmdq_tx_timeout(opcode, priv->cmq.tx_timeout);
 	u32 timeout = 0;
 
 	do {
@@ -1299,8 +1297,9 @@ static void hns_roce_wait_csq_done(struct hns_roce_dev *hr_dev, u16 opcode)
 	} while (++timeout < tx_timeout);
 }
 
-static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
-			       struct hns_roce_cmq_desc *desc, int num)
+static int __hns_roce_cmq_send_one(struct hns_roce_dev *hr_dev,
+				   struct hns_roce_cmq_desc *desc,
+				   int num, u32 tx_timeout)
 {
 	struct hns_roce_v2_priv *priv = hr_dev->priv;
 	struct hns_roce_v2_cmq_ring *csq = &priv->cmq.csq;
@@ -1309,8 +1308,6 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 	int ret;
 	int i;
 
-	spin_lock_bh(&csq->lock);
-
 	tail = csq->head;
 
 	for (i = 0; i < num; i++) {
@@ -1324,22 +1321,17 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 
 	atomic64_inc(&hr_dev->dfx_cnt[HNS_ROCE_DFX_CMDS_CNT]);
 
-	hns_roce_wait_csq_done(hr_dev, le16_to_cpu(desc->opcode));
+	hns_roce_wait_csq_done(hr_dev, tx_timeout);
 	if (hns_roce_cmq_csq_done(hr_dev)) {
 		ret = 0;
 		for (i = 0; i < num; i++) {
 			/* check the result of hardware write back */
-			desc[i] = csq->desc[tail++];
+			desc_ret = le16_to_cpu(csq->desc[tail++].retval);
 			if (tail == csq->desc_num)
 				tail = 0;
-
-			desc_ret = le16_to_cpu(desc[i].retval);
 			if (likely(desc_ret == CMD_EXEC_SUCCESS))
 				continue;
 
-			dev_err_ratelimited(hr_dev->dev,
-					    "Cmdq IO error, opcode = 0x%x, return = 0x%x.\n",
-					    desc->opcode, desc_ret);
 			ret = hns_roce_cmd_err_convert_errno(desc_ret);
 		}
 	} else {
@@ -1354,14 +1346,54 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 		ret = -EAGAIN;
 	}
 
-	spin_unlock_bh(&csq->lock);
-
 	if (ret)
 		atomic64_inc(&hr_dev->dfx_cnt[HNS_ROCE_DFX_CMDS_ERR_CNT]);
 
 	return ret;
 }
 
+static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
+			       struct hns_roce_cmq_desc *desc, int num)
+{
+	struct hns_roce_v2_priv *priv = hr_dev->priv;
+	struct hns_roce_v2_cmq_ring *csq = &priv->cmq.csq;
+	u16 opcode = le16_to_cpu(desc->opcode);
+	u32 tx_timeout = hns_roce_cmdq_tx_timeout(opcode, priv->cmq.tx_timeout);
+	u8 try_cnt = HNS_ROCE_OPC_POST_MB_TRY_CNT;
+	u32 rsv_tail;
+	int ret;
+	int i;
+
+	while (try_cnt) {
+		try_cnt--;
+
+		spin_lock_bh(&csq->lock);
+		rsv_tail = csq->head;
+		ret = __hns_roce_cmq_send_one(hr_dev, desc, num, tx_timeout);
+		if (opcode == HNS_ROCE_OPC_POST_MB && ret == -ETIME &&
+		    try_cnt) {
+			spin_unlock_bh(&csq->lock);
+			mdelay(HNS_ROCE_OPC_POST_MB_RETRY_GAP_MSEC);
+			continue;
+		}
+
+		for (i = 0; i < num; i++) {
+			desc[i] = csq->desc[rsv_tail++];
+			if (rsv_tail == csq->desc_num)
+				rsv_tail = 0;
+		}
+		spin_unlock_bh(&csq->lock);
+		break;
+	}
+
+	if (ret)
+		dev_err_ratelimited(hr_dev->dev,
+				    "Cmdq IO error, opcode = 0x%x, return = %d.\n",
+				    opcode, ret);
+
+	return ret;
+}
+
 static int hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
 			     struct hns_roce_cmq_desc *desc, int num)
 {
diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
index cbdbc9edbce6..91a5665465ff 100644
--- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
+++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
@@ -230,6 +230,8 @@ enum hns_roce_opcode_type {
 };
 
 #define HNS_ROCE_OPC_POST_MB_TIMEOUT 35000
+#define HNS_ROCE_OPC_POST_MB_TRY_CNT 8
+#define HNS_ROCE_OPC_POST_MB_RETRY_GAP_MSEC 5
 struct hns_roce_cmdq_tx_timeout_map {
 	u16 opcode;
 	u32 tx_timeout;
diff --git a/drivers/infiniband/hw/mana/main.c b/drivers/infiniband/hw/mana/main.c
index 67c2d43135a8..457cea6d9909 100644
--- a/drivers/infiniband/hw/mana/main.c
+++ b/drivers/infiniband/hw/mana/main.c
@@ -174,7 +174,7 @@ static int mana_gd_allocate_doorbell_page(struct gdma_context *gc,
 
 	req.resource_type = GDMA_RESOURCE_DOORBELL_PAGE;
 	req.num_resources = 1;
-	req.alignment = 1;
+	req.alignment = PAGE_SIZE / MANA_PAGE_SIZE;
 
 	/* Have GDMA start searching from 0 */
 	req.allocated_resources = 0;
diff --git a/drivers/infiniband/hw/mlx5/ah.c b/drivers/infiniband/hw/mlx5/ah.c
index 505bc47fd575..99036afb3aef 100644
--- a/drivers/infiniband/hw/mlx5/ah.c
+++ b/drivers/infiniband/hw/mlx5/ah.c
@@ -67,7 +67,8 @@ static void create_ib_ah(struct mlx5_ib_dev *dev, struct mlx5_ib_ah *ah,
 		ah->av.tclass = grh->traffic_class;
 	}
 
-	ah->av.stat_rate_sl = (rdma_ah_get_static_rate(ah_attr) << 4);
+	ah->av.stat_rate_sl =
+		(mlx5r_ib_rate(dev, rdma_ah_get_static_rate(ah_attr)) << 4);
 
 	if (ah_attr->type == RDMA_AH_ATTR_TYPE_ROCE) {
 		if (init_attr->xmit_slave)
diff --git a/drivers/infiniband/hw/mlx5/counters.c b/drivers/infiniband/hw/mlx5/counters.c
index 4f6c1968a2ee..81cfa74147a1 100644
--- a/drivers/infiniband/hw/mlx5/counters.c
+++ b/drivers/infiniband/hw/mlx5/counters.c
@@ -546,6 +546,7 @@ static int mlx5_ib_counter_bind_qp(struct rdma_counter *counter,
 				   struct ib_qp *qp)
 {
 	struct mlx5_ib_dev *dev = to_mdev(qp->device);
+	bool new = false;
 	int err;
 
 	if (!counter->id) {
@@ -560,6 +561,7 @@ static int mlx5_ib_counter_bind_qp(struct rdma_counter *counter,
 			return err;
 		counter->id =
 			MLX5_GET(alloc_q_counter_out, out, counter_set_id);
+		new = true;
 	}
 
 	err = mlx5_ib_qp_set_counter(qp, counter);
@@ -569,8 +571,10 @@ static int mlx5_ib_counter_bind_qp(struct rdma_counter *counter,
 	return 0;
 
 fail_set_counter:
-	mlx5_ib_counter_dealloc(counter);
-	counter->id = 0;
+	if (new) {
+		mlx5_ib_counter_dealloc(counter);
+		counter->id = 0;
+	}
 
 	return err;
 }
diff --git a/drivers/infiniband/hw/mlx5/mr.c b/drivers/infiniband/hw/mlx5/mr.c
index bb02b6adbf2c..753faa9ad06a 100644
--- a/drivers/infiniband/hw/mlx5/mr.c
+++ b/drivers/infiniband/hw/mlx5/mr.c
@@ -1550,7 +1550,7 @@ static void mlx5_ib_dmabuf_invalidate_cb(struct dma_buf_attachment *attach)
 
 	dma_resv_assert_held(umem_dmabuf->attach->dmabuf->resv);
 
-	if (!umem_dmabuf->sgt)
+	if (!umem_dmabuf->sgt || !mr)
 		return;
 
 	mlx5r_umr_update_mr_pas(mr, MLX5_IB_UPD_XLT_ZAP);
@@ -1935,7 +1935,8 @@ err:
 static void
 mlx5_free_priv_descs(struct mlx5_ib_mr *mr)
 {
-	if (!mr->umem && !mr->data_direct && mr->descs) {
+	if (!mr->umem && !mr->data_direct &&
+	    mr->ibmr.type != IB_MR_TYPE_DM && mr->descs) {
 		struct ib_device *device = mr->ibmr.device;
 		int size = mr->max_descs * mr->desc_size;
 		struct mlx5_ib_dev *dev = to_mdev(device);
@@ -2022,11 +2023,16 @@ static int mlx5_revoke_mr(struct mlx5_ib_mr *mr)
 	struct mlx5_ib_dev *dev = to_mdev(mr->ibmr.device);
 	struct mlx5_cache_ent *ent = mr->mmkey.cache_ent;
 	bool is_odp = is_odp_mr(mr);
+	bool is_odp_dma_buf = is_dmabuf_mr(mr) &&
+			!to_ib_umem_dmabuf(mr->umem)->pinned;
 	int ret = 0;
 
 	if (is_odp)
 		mutex_lock(&to_ib_umem_odp(mr->umem)->umem_mutex);
 
+	if (is_odp_dma_buf)
+		dma_resv_lock(to_ib_umem_dmabuf(mr->umem)->attach->dmabuf->resv, NULL);
+
 	if (mr->mmkey.cacheable && !mlx5r_umr_revoke_mr(mr) && !cache_ent_find_and_store(dev, mr)) {
 		ent = mr->mmkey.cache_ent;
 		/* upon storing to a clean temp entry - schedule its cleanup */
@@ -2054,6 +2060,12 @@ out:
 		mutex_unlock(&to_ib_umem_odp(mr->umem)->umem_mutex);
 	}
 
+	if (is_odp_dma_buf) {
+		if (!ret)
+			to_ib_umem_dmabuf(mr->umem)->private = NULL;
+		dma_resv_unlock(to_ib_umem_dmabuf(mr->umem)->attach->dmabuf->resv);
+	}
+
 	return ret;
 }
 
diff --git a/drivers/infiniband/hw/mlx5/odp.c b/drivers/infiniband/hw/mlx5/odp.c
index f1e23583e6c0..e77c9280c07e 100644
--- a/drivers/infiniband/hw/mlx5/odp.c
+++ b/drivers/infiniband/hw/mlx5/odp.c
@@ -242,6 +242,7 @@ static void destroy_unused_implicit_child_mr(struct mlx5_ib_mr *mr)
 	if (__xa_cmpxchg(&imr->implicit_children, idx, mr, NULL, GFP_KERNEL) !=
 	    mr) {
 		xa_unlock(&imr->implicit_children);
+		mlx5r_deref_odp_mkey(&imr->mmkey);
 		return;
 	}
 
diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c
index a43eba9d3572..88724d15705d 100644
--- a/drivers/infiniband/hw/mlx5/qp.c
+++ b/drivers/infiniband/hw/mlx5/qp.c
@@ -3447,11 +3447,11 @@ static int ib_to_mlx5_rate_map(u8 rate)
 	return 0;
 }
 
-static int ib_rate_to_mlx5(struct mlx5_ib_dev *dev, u8 rate)
+int mlx5r_ib_rate(struct mlx5_ib_dev *dev, u8 rate)
 {
 	u32 stat_rate_support;
 
-	if (rate == IB_RATE_PORT_CURRENT)
+	if (rate == IB_RATE_PORT_CURRENT || rate == IB_RATE_800_GBPS)
 		return 0;
 
 	if (rate < IB_RATE_2_5_GBPS || rate > IB_RATE_800_GBPS)
@@ -3596,7 +3596,7 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp,
 		       sizeof(grh->dgid.raw));
 	}
 
-	err = ib_rate_to_mlx5(dev, rdma_ah_get_static_rate(ah));
+	err = mlx5r_ib_rate(dev, rdma_ah_get_static_rate(ah));
 	if (err < 0)
 		return err;
 	MLX5_SET(ads, path, stat_rate, err);
@@ -4579,6 +4579,8 @@ static int mlx5_ib_modify_dct(struct ib_qp *ibqp, struct ib_qp_attr *attr,
 
 		set_id = mlx5_ib_get_counters_id(dev, attr->port_num - 1);
 		MLX5_SET(dctc, dctc, counter_set_id, set_id);
+
+		qp->port = attr->port_num;
 	} else if (cur_state == IB_QPS_INIT && new_state == IB_QPS_RTR) {
 		struct mlx5_ib_modify_qp_resp resp = {};
 		u32 out[MLX5_ST_SZ_DW(create_dct_out)] = {};
@@ -5074,7 +5076,7 @@ static int mlx5_ib_dct_query_qp(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *mqp,
 	}
 
 	if (qp_attr_mask & IB_QP_PORT)
-		qp_attr->port_num = MLX5_GET(dctc, dctc, port);
+		qp_attr->port_num = mqp->port;
 	if (qp_attr_mask & IB_QP_MIN_RNR_TIMER)
 		qp_attr->min_rnr_timer = MLX5_GET(dctc, dctc, min_rnr_nak);
 	if (qp_attr_mask & IB_QP_AV) {
diff --git a/drivers/infiniband/hw/mlx5/qp.h b/drivers/infiniband/hw/mlx5/qp.h
index b6ee7c3ee1ca..2530e7730635 100644
--- a/drivers/infiniband/hw/mlx5/qp.h
+++ b/drivers/infiniband/hw/mlx5/qp.h
@@ -56,4 +56,5 @@ int mlx5_core_xrcd_dealloc(struct mlx5_ib_dev *dev, u32 xrcdn);
 int mlx5_ib_qp_set_counter(struct ib_qp *qp, struct rdma_counter *counter);
 int mlx5_ib_qp_event_init(void);
 void mlx5_ib_qp_event_cleanup(void);
+int mlx5r_ib_rate(struct mlx5_ib_dev *dev, u8 rate);
 #endif /* _MLX5_IB_QP_H */
diff --git a/drivers/infiniband/hw/mlx5/umr.c b/drivers/infiniband/hw/mlx5/umr.c
index 887fd6fa3ba9..793f3c5c4d01 100644
--- a/drivers/infiniband/hw/mlx5/umr.c
+++ b/drivers/infiniband/hw/mlx5/umr.c
@@ -231,30 +231,6 @@ void mlx5r_umr_cleanup(struct mlx5_ib_dev *dev)
 	ib_dealloc_pd(dev->umrc.pd);
 }
 
-static int mlx5r_umr_recover(struct mlx5_ib_dev *dev)
-{
-	struct umr_common *umrc = &dev->umrc;
-	struct ib_qp_attr attr;
-	int err;
-
-	attr.qp_state = IB_QPS_RESET;
-	err = ib_modify_qp(umrc->qp, &attr, IB_QP_STATE);
-	if (err) {
-		mlx5_ib_dbg(dev, "Couldn't modify UMR QP\n");
-		goto err;
-	}
-
-	err = mlx5r_umr_qp_rst2rts(dev, umrc->qp);
-	if (err)
-		goto err;
-
-	umrc->state = MLX5_UMR_STATE_ACTIVE;
-	return 0;
-
-err:
-	umrc->state = MLX5_UMR_STATE_ERR;
-	return err;
-}
 
 static int mlx5r_umr_post_send(struct ib_qp *ibqp, u32 mkey, struct ib_cqe *cqe,
 			       struct mlx5r_umr_wqe *wqe, bool with_data)
@@ -302,6 +278,61 @@ out:
 	return err;
 }
 
+static int mlx5r_umr_recover(struct mlx5_ib_dev *dev, u32 mkey,
+			     struct mlx5r_umr_context *umr_context,
+			     struct mlx5r_umr_wqe *wqe, bool with_data)
+{
+	struct umr_common *umrc = &dev->umrc;
+	struct ib_qp_attr attr;
+	int err;
+
+	mutex_lock(&umrc->lock);
+	/* Preventing any further WRs to be sent now */
+	if (umrc->state != MLX5_UMR_STATE_RECOVER) {
+		mlx5_ib_warn(dev, "UMR recovery encountered an unexpected state=%d\n",
+			     umrc->state);
+		umrc->state = MLX5_UMR_STATE_RECOVER;
+	}
+	mutex_unlock(&umrc->lock);
+
+	/* Sending a final/barrier WR (the failed one) and wait for its completion.
+	 * This will ensure that all the previous WRs got a completion before
+	 * we set the QP state to RESET.
+	 */
+	err = mlx5r_umr_post_send(umrc->qp, mkey, &umr_context->cqe, wqe,
+				  with_data);
+	if (err) {
+		mlx5_ib_warn(dev, "UMR recovery post send failed, err %d\n", err);
+		goto err;
+	}
+
+	/* Since the QP is in an error state, it will only receive
+	 * IB_WC_WR_FLUSH_ERR. However, as it serves only as a barrier
+	 * we don't care about its status.
+	 */
+	wait_for_completion(&umr_context->done);
+
+	attr.qp_state = IB_QPS_RESET;
+	err = ib_modify_qp(umrc->qp, &attr, IB_QP_STATE);
+	if (err) {
+		mlx5_ib_warn(dev, "Couldn't modify UMR QP to RESET, err=%d\n", err);
+		goto err;
+	}
+
+	err = mlx5r_umr_qp_rst2rts(dev, umrc->qp);
+	if (err) {
+		mlx5_ib_warn(dev, "Couldn't modify UMR QP to RTS, err=%d\n", err);
+		goto err;
+	}
+
+	umrc->state = MLX5_UMR_STATE_ACTIVE;
+	return 0;
+
+err:
+	umrc->state = MLX5_UMR_STATE_ERR;
+	return err;
+}
+
 static void mlx5r_umr_done(struct ib_cq *cq, struct ib_wc *wc)
 {
 	struct mlx5_ib_umr_context *context =
@@ -366,9 +397,7 @@ static int mlx5r_umr_post_send_wait(struct mlx5_ib_dev *dev, u32 mkey,
 		mlx5_ib_warn(dev,
 			"reg umr failed (%u). Trying to recover and resubmit the flushed WQEs, mkey = %u\n",
 			umr_context.status, mkey);
-		mutex_lock(&umrc->lock);
-		err = mlx5r_umr_recover(dev);
-		mutex_unlock(&umrc->lock);
+		err = mlx5r_umr_recover(dev, mkey, &umr_context, wqe, with_data);
 		if (err)
 			mlx5_ib_warn(dev, "couldn't recover UMR, err %d\n",
 				     err);