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

Pull rdma fixes from Jason Gunthorpe:
 "Several mlx5 bugs, crashers, and reports:

   - Limit stack usage

   - Fix mis-use of __xa_store/erase() without holding the lock to a
     locked version

   - Rate limit prints in the gid cache error cases

   - Fully initialize the event object before making it globally visible
     in an xarray

   - Fix deadlock inside the ODP code if the MMU notifier was called
     from a reclaim context

   - Include missed counters for some switchdev configurations and
     mulit-port MPV mode

   - Fix loopback packet support when in mulit-port MPV mode"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/mlx5: Fix vport loopback for MPV device
  RDMA/mlx5: Fix CC counters query for MPV
  RDMA/mlx5: Fix HW counters query for non-representor devices
  IB/core: Annotate umem_mutex acquisition under fs_reclaim for lockdep
  IB/mlx5: Fix potential deadlock in MR deregistration
  RDMA/mlx5: Initialize obj_event->obj_sub_list before xa_insert
  RDMA/core: Rate limit GID cache warning messages
  RDMA/mlx5: Fix unsafe xarray access in implicit ODP handling
  RDMA/mlx5: reduce stack usage in mlx5_ib_ufile_hw_cleanup

drivers/infiniband/core/cache.c

@@ -582,8 +582,8 @@ static int __ib_cache_gid_add(struct ib_device *ib_dev, u32 port,
 out_unlock:
 	mutex_unlock(&table->lock);
 	if (ret)
-		pr_warn("%s: unable to add gid %pI6 error=%d\n",
-			__func__, gid->raw, ret);
+		pr_warn_ratelimited("%s: unable to add gid %pI6 error=%d\n",
+				    __func__, gid->raw, ret);
 	return ret;
 }
 

drivers/infiniband/core/umem_odp.c

@@ -76,6 +76,17 @@ static int ib_init_umem_odp(struct ib_umem_odp *umem_odp,
 	end = ALIGN(end, page_size);
 	if (unlikely(end < page_size))
 		return -EOVERFLOW;
+	/*
+	 * The mmu notifier can be called within reclaim contexts and takes the
+	 * umem_mutex. This is rare to trigger in testing, teach lockdep about
+	 * it.
+	 */
+	if (IS_ENABLED(CONFIG_LOCKDEP)) {
+		fs_reclaim_acquire(GFP_KERNEL);
+		mutex_lock(&umem_odp->umem_mutex);
+		mutex_unlock(&umem_odp->umem_mutex);
+		fs_reclaim_release(GFP_KERNEL);
+	}
 
 	nr_entries = (end - start) >> PAGE_SHIFT;
 	if (!(nr_entries * PAGE_SIZE / page_size))

drivers/infiniband/hw/mlx5/counters.c

@@ -398,7 +398,7 @@ static int do_get_hw_stats(struct ib_device *ibdev,
 		return ret;
 
 	/* We don't expose device counters over Vports */
-	if (is_mdev_switchdev_mode(dev->mdev) && port_num != 0)
+	if (is_mdev_switchdev_mode(dev->mdev) && dev->is_rep && port_num != 0)
 		goto done;
 
 	if (MLX5_CAP_PCAM_FEATURE(dev->mdev, rx_icrc_encapsulated_counter)) {
@@ -418,7 +418,7 @@ static int do_get_hw_stats(struct ib_device *ibdev,
 			 */
 			goto done;
 		}
-		ret = mlx5_lag_query_cong_counters(dev->mdev,
+		ret = mlx5_lag_query_cong_counters(mdev,
 						   stats->value +
 						   cnts->num_q_counters,
 						   cnts->num_cong_counters,

drivers/infiniband/hw/mlx5/devx.c

@@ -1958,6 +1958,7 @@ subscribe_event_xa_alloc(struct mlx5_devx_event_table *devx_event_table,
 			/* Level1 is valid for future use, no need to free */
 			return -ENOMEM;
 
+		INIT_LIST_HEAD(&obj_event->obj_sub_list);
 		err = xa_insert(&event->object_ids,
 				key_level2,
 				obj_event,
@@ -1966,7 +1967,6 @@ subscribe_event_xa_alloc(struct mlx5_devx_event_table *devx_event_table,
 			kfree(obj_event);
 			return err;
 		}
-		INIT_LIST_HEAD(&obj_event->obj_sub_list);
 	}
 
 	return 0;
@@ -2669,7 +2669,7 @@ static void devx_wait_async_destroy(struct mlx5_async_cmd *cmd)
 
 void mlx5_ib_ufile_hw_cleanup(struct ib_uverbs_file *ufile)
 {
-	struct mlx5_async_cmd async_cmd[MAX_ASYNC_CMDS];
+	struct mlx5_async_cmd *async_cmd;
 	struct ib_ucontext *ucontext = ufile->ucontext;
 	struct ib_device *device = ucontext->device;
 	struct mlx5_ib_dev *dev = to_mdev(device);
@@ -2678,6 +2678,10 @@ void mlx5_ib_ufile_hw_cleanup(struct ib_uverbs_file *ufile)
 	int head = 0;
 	int tail = 0;
 
+	async_cmd = kcalloc(MAX_ASYNC_CMDS, sizeof(*async_cmd), GFP_KERNEL);
+	if (!async_cmd)
+		return;
+
 	list_for_each_entry(uobject, &ufile->uobjects, list) {
 		WARN_ON(uverbs_try_lock_object(uobject, UVERBS_LOOKUP_WRITE));
 
@@ -2713,6 +2717,8 @@ void mlx5_ib_ufile_hw_cleanup(struct ib_uverbs_file *ufile)
 		devx_wait_async_destroy(&async_cmd[head % MAX_ASYNC_CMDS]);
 		head++;
 	}
+
+	kfree(async_cmd);
 }
 
 static ssize_t devx_async_cmd_event_read(struct file *filp, char __user *buf,

drivers/infiniband/hw/mlx5/main.c

@@ -1791,6 +1791,33 @@ static void deallocate_uars(struct mlx5_ib_dev *dev,
 					     context->devx_uid);
 }
 
+static int mlx5_ib_enable_lb_mp(struct mlx5_core_dev *master,
+				struct mlx5_core_dev *slave)
+{
+	int err;
+
+	err = mlx5_nic_vport_update_local_lb(master, true);
+	if (err)
+		return err;
+
+	err = mlx5_nic_vport_update_local_lb(slave, true);
+	if (err)
+		goto out;
+
+	return 0;
+
+out:
+	mlx5_nic_vport_update_local_lb(master, false);
+	return err;
+}
+
+static void mlx5_ib_disable_lb_mp(struct mlx5_core_dev *master,
+				  struct mlx5_core_dev *slave)
+{
+	mlx5_nic_vport_update_local_lb(slave, false);
+	mlx5_nic_vport_update_local_lb(master, false);
+}
+
 int mlx5_ib_enable_lb(struct mlx5_ib_dev *dev, bool td, bool qp)
 {
 	int err = 0;
@@ -3495,6 +3522,8 @@ static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
 
 	lockdep_assert_held(&mlx5_ib_multiport_mutex);
 
+	mlx5_ib_disable_lb_mp(ibdev->mdev, mpi->mdev);
+
 	mlx5_core_mp_event_replay(ibdev->mdev,
 				  MLX5_DRIVER_EVENT_AFFILIATION_REMOVED,
 				  NULL);
@@ -3590,6 +3619,10 @@ static bool mlx5_ib_bind_slave_port(struct mlx5_ib_dev *ibdev,
 				  MLX5_DRIVER_EVENT_AFFILIATION_DONE,
 				  &key);
 
+	err = mlx5_ib_enable_lb_mp(ibdev->mdev, mpi->mdev);
+	if (err)
+		goto unbind;
+
 	return true;
 
 unbind:

drivers/infiniband/hw/mlx5/mr.c

@@ -2027,23 +2027,50 @@ void mlx5_ib_revoke_data_direct_mrs(struct mlx5_ib_dev *dev)
 	}
 }
 
-static int mlx5_revoke_mr(struct mlx5_ib_mr *mr)
+static int mlx5_umr_revoke_mr_with_lock(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;
-	bool from_cache = !!ent;
-	int ret = 0;
+			      !to_ib_umem_dmabuf(mr->umem)->pinned;
+	bool is_odp = is_odp_mr(mr);
+	int ret;
 
 	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);
+		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)) {
+	ret = mlx5r_umr_revoke_mr(mr);
+
+	if (is_odp) {
+		if (!ret)
+			to_ib_umem_odp(mr->umem)->private = NULL;
+		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;
+}
+
+static int mlx5r_handle_mkey_cleanup(struct mlx5_ib_mr *mr)
+{
+	bool is_odp_dma_buf = is_dmabuf_mr(mr) &&
+			      !to_ib_umem_dmabuf(mr->umem)->pinned;
+	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 from_cache = !!ent;
+	int ret;
+
+	if (mr->mmkey.cacheable && !mlx5_umr_revoke_mr_with_lock(mr) &&
+	    !cache_ent_find_and_store(dev, mr)) {
 		ent = mr->mmkey.cache_ent;
 		/* upon storing to a clean temp entry - schedule its cleanup */
 		spin_lock_irq(&ent->mkeys_queue.lock);
@@ -2055,7 +2082,7 @@ static int mlx5_revoke_mr(struct mlx5_ib_mr *mr)
 			ent->tmp_cleanup_scheduled = true;
 		}
 		spin_unlock_irq(&ent->mkeys_queue.lock);
-		goto out;
+		return 0;
 	}
 
 	if (ent) {
@@ -2064,8 +2091,14 @@ static int mlx5_revoke_mr(struct mlx5_ib_mr *mr)
 		mr->mmkey.cache_ent = NULL;
 		spin_unlock_irq(&ent->mkeys_queue.lock);
 	}
+
+	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);
 	ret = destroy_mkey(dev, mr);
-out:
 	if (is_odp) {
 		if (!ret)
 			to_ib_umem_odp(mr->umem)->private = NULL;
@@ -2075,9 +2108,9 @@ out:
 	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);
+		dma_resv_unlock(
+			to_ib_umem_dmabuf(mr->umem)->attach->dmabuf->resv);
 	}
-
 	return ret;
 }
 
@@ -2126,7 +2159,7 @@ static int __mlx5_ib_dereg_mr(struct ib_mr *ibmr)
 	}
 
 	/* Stop DMA */
-	rc = mlx5_revoke_mr(mr);
+	rc = mlx5r_handle_mkey_cleanup(mr);
 	if (rc)
 		return rc;
 

drivers/infiniband/hw/mlx5/odp.c

@@ -259,8 +259,8 @@ static void destroy_unused_implicit_child_mr(struct mlx5_ib_mr *mr)
 	}
 
 	if (MLX5_CAP_ODP(mr_to_mdev(mr)->mdev, mem_page_fault))
-		__xa_erase(&mr_to_mdev(mr)->odp_mkeys,
-			   mlx5_base_mkey(mr->mmkey.key));
+		xa_erase(&mr_to_mdev(mr)->odp_mkeys,
+			 mlx5_base_mkey(mr->mmkey.key));
 	xa_unlock(&imr->implicit_children);
 
 	/* Freeing a MR is a sleeping operation, so bounce to a work queue */
@@ -532,8 +532,8 @@ static struct mlx5_ib_mr *implicit_get_child_mr(struct mlx5_ib_mr *imr,
 	}
 
 	if (MLX5_CAP_ODP(dev->mdev, mem_page_fault)) {
-		ret = __xa_store(&dev->odp_mkeys, mlx5_base_mkey(mr->mmkey.key),
-				 &mr->mmkey, GFP_KERNEL);
+		ret = xa_store(&dev->odp_mkeys, mlx5_base_mkey(mr->mmkey.key),
+			       &mr->mmkey, GFP_KERNEL);
 		if (xa_is_err(ret)) {
 			ret = ERR_PTR(xa_err(ret));
 			__xa_erase(&imr->implicit_children, idx);