1 files changed, 79 insertions, 99 deletions

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 831539419c30..51100350b688 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -67,6 +67,7 @@
 #include <rdma/uverbs_std_types.h>
 #include <rdma/mlx5_user_ioctl_verbs.h>
 #include <rdma/mlx5_user_ioctl_cmds.h>
+#include <rdma/ib_umem_odp.h>
 
 #define UVERBS_MODULE_NAME mlx5_ib
 #include <rdma/uverbs_named_ioctl.h>
@@ -693,21 +694,6 @@ static void get_atomic_caps_qp(struct mlx5_ib_dev *dev,
 	get_atomic_caps(dev, atomic_size_qp, props);
 }
 
-static void get_atomic_caps_dc(struct mlx5_ib_dev *dev,
-			       struct ib_device_attr *props)
-{
-	u8 atomic_size_qp = MLX5_CAP_ATOMIC(dev->mdev, atomic_size_dc);
-
-	get_atomic_caps(dev, atomic_size_qp, props);
-}
-
-bool mlx5_ib_dc_atomic_is_supported(struct mlx5_ib_dev *dev)
-{
-	struct ib_device_attr props = {};
-
-	get_atomic_caps_dc(dev, &props);
-	return (props.atomic_cap == IB_ATOMIC_HCA) ? true : false;
-}
 static int mlx5_query_system_image_guid(struct ib_device *ibdev,
 					__be64 *sys_image_guid)
 {
@@ -844,8 +830,8 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	resp_len = sizeof(resp.comp_mask) + sizeof(resp.response_length);
 	if (uhw->outlen && uhw->outlen < resp_len)
 		return -EINVAL;
-	else
-		resp.response_length = resp_len;
+
+	resp.response_length = resp_len;
 
 	if (uhw->inlen && !ib_is_udata_cleared(uhw, 0, uhw->inlen))
 		return -EINVAL;
@@ -1011,6 +997,8 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 		1 << MLX5_CAP_GEN(mdev, log_max_klm_list_size);
 	props->max_pi_fast_reg_page_list_len =
 		props->max_fast_reg_page_list_len / 2;
+	props->max_sgl_rd =
+		MLX5_CAP_GEN(mdev, max_sgl_for_optimized_performance);
 	get_atomic_caps_qp(dev, props);
 	props->masked_atomic_cap   = IB_ATOMIC_NONE;
 	props->max_mcast_grp	   = 1 << MLX5_CAP_GEN(mdev, log_max_mcg);
@@ -1031,7 +1019,7 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	if (MLX5_CAP_GEN(mdev, cd))
 		props->device_cap_flags |= IB_DEVICE_CROSS_CHANNEL;
 
-	if (!mlx5_core_is_pf(mdev))
+	if (mlx5_core_is_vf(mdev))
 		props->device_cap_flags |= IB_DEVICE_VIRTUAL_FUNCTION;
 
 	if (mlx5_ib_port_link_layer(ibdev, 1) ==
@@ -1161,8 +1149,14 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 				MLX5_MIN_SINGLE_STRIDE_LOG_NUM_BYTES;
 			resp.striding_rq_caps.max_single_stride_log_num_of_bytes =
 				MLX5_MAX_SINGLE_STRIDE_LOG_NUM_BYTES;
-			resp.striding_rq_caps.min_single_wqe_log_num_of_strides =
-				MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES;
+			if (MLX5_CAP_GEN(dev->mdev, ext_stride_num_range))
+				resp.striding_rq_caps
+					.min_single_wqe_log_num_of_strides =
+					MLX5_EXT_MIN_SINGLE_WQE_LOG_NUM_STRIDES;
+			else
+				resp.striding_rq_caps
+					.min_single_wqe_log_num_of_strides =
+					MLX5_MIN_SINGLE_WQE_LOG_NUM_STRIDES;
 			resp.striding_rq_caps.max_single_wqe_log_num_of_strides =
 				MLX5_MAX_SINGLE_WQE_LOG_NUM_STRIDES;
 			resp.striding_rq_caps.supported_qpts =
@@ -1808,7 +1802,7 @@ static int mlx5_ib_alloc_ucontext(struct ib_ucontext *uctx,
 		return -EINVAL;
 
 	resp.qp_tab_size = 1 << MLX5_CAP_GEN(dev->mdev, log_max_qp);
-	if (mlx5_core_is_pf(dev->mdev) && MLX5_CAP_GEN(dev->mdev, bf))
+	if (dev->wc_support)
 		resp.bf_reg_size = 1 << MLX5_CAP_GEN(dev->mdev, log_bf_reg_size);
 	resp.cache_line_size = cache_line_size();
 	resp.max_sq_desc_sz = MLX5_CAP_GEN(dev->mdev, max_wqe_sz_sq);
@@ -2168,7 +2162,7 @@ static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd,
 	mlx5_ib_dbg(dev, "uar idx 0x%lx, pfn %pa\n", idx, &pfn);
 
 	err = rdma_user_mmap_io(&context->ibucontext, vma, pfn, PAGE_SIZE,
-				prot);
+				prot, NULL);
 	if (err) {
 		mlx5_ib_err(dev,
 			    "rdma_user_mmap_io failed with error=%d, mmap_cmd=%s\n",
@@ -2210,7 +2204,8 @@ static int dm_mmap(struct ib_ucontext *context, struct vm_area_struct *vma)
 	      PAGE_SHIFT) +
 	      page_idx;
 	return rdma_user_mmap_io(context, vma, pfn, map_size,
-				 pgprot_writecombine(vma->vm_page_prot));
+				 pgprot_writecombine(vma->vm_page_prot),
+				 NULL);
 }
 
 static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma)
@@ -2248,7 +2243,8 @@ static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vm
 			PAGE_SHIFT;
 		return rdma_user_mmap_io(&context->ibucontext, vma, pfn,
 					 PAGE_SIZE,
-					 pgprot_noncached(vma->vm_page_prot));
+					 pgprot_noncached(vma->vm_page_prot),
+					 NULL);
 	case MLX5_IB_MMAP_CLOCK_INFO:
 		return mlx5_ib_mmap_clock_info_page(dev, vma, context);
 
@@ -5145,8 +5141,7 @@ static int mlx5_port_immutable(struct ib_device *ibdev, u8 port_num,
 	immutable->pkey_tbl_len = attr.pkey_tbl_len;
 	immutable->gid_tbl_len = attr.gid_tbl_len;
 	immutable->core_cap_flags = get_core_cap_flags(ibdev, &rep);
-	if ((ll == IB_LINK_LAYER_INFINIBAND) || MLX5_CAP_GEN(dev->mdev, roce))
-		immutable->max_mad_size = IB_MGMT_MAD_SIZE;
+	immutable->max_mad_size = IB_MGMT_MAD_SIZE;
 
 	return 0;
 }
@@ -5249,11 +5244,9 @@ static int mlx5_enable_eth(struct mlx5_ib_dev *dev)
 {
 	int err;
 
-	if (MLX5_CAP_GEN(dev->mdev, roce)) {
-		err = mlx5_nic_vport_enable_roce(dev->mdev);
-		if (err)
-			return err;
-	}
+	err = mlx5_nic_vport_enable_roce(dev->mdev);
+	if (err)
+		return err;
 
 	err = mlx5_eth_lag_init(dev);
 	if (err)
@@ -5262,8 +5255,7 @@ static int mlx5_enable_eth(struct mlx5_ib_dev *dev)
 	return 0;
 
 err_disable_roce:
-	if (MLX5_CAP_GEN(dev->mdev, roce))
-		mlx5_nic_vport_disable_roce(dev->mdev);
+	mlx5_nic_vport_disable_roce(dev->mdev);
 
 	return err;
 }
@@ -5271,8 +5263,7 @@ err_disable_roce:
 static void mlx5_disable_eth(struct mlx5_ib_dev *dev)
 {
 	mlx5_eth_lag_cleanup(dev);
-	if (MLX5_CAP_GEN(dev->mdev, roce))
-		mlx5_nic_vport_disable_roce(dev->mdev);
+	mlx5_nic_vport_disable_roce(dev->mdev);
 }
 
 struct mlx5_ib_counter {
@@ -5710,11 +5701,10 @@ static int mlx5_ib_rn_get_params(struct ib_device *device, u8 port_num,
 
 static void delay_drop_debugfs_cleanup(struct mlx5_ib_dev *dev)
 {
-	if (!dev->delay_drop.dbg)
+	if (!dev->delay_drop.dir_debugfs)
 		return;
-	debugfs_remove_recursive(dev->delay_drop.dbg->dir_debugfs);
-	kfree(dev->delay_drop.dbg);
-	dev->delay_drop.dbg = NULL;
+	debugfs_remove_recursive(dev->delay_drop.dir_debugfs);
+	dev->delay_drop.dir_debugfs = NULL;
 }
 
 static void cancel_delay_drop(struct mlx5_ib_dev *dev)
@@ -5765,52 +5755,22 @@ static const struct file_operations fops_delay_drop_timeout = {
 	.read	= delay_drop_timeout_read,
 };
 
-static int delay_drop_debugfs_init(struct mlx5_ib_dev *dev)
+static void delay_drop_debugfs_init(struct mlx5_ib_dev *dev)
 {
-	struct mlx5_ib_dbg_delay_drop *dbg;
+	struct dentry *root;
 
 	if (!mlx5_debugfs_root)
-		return 0;
-
-	dbg = kzalloc(sizeof(*dbg), GFP_KERNEL);
-	if (!dbg)
-		return -ENOMEM;
-
-	dev->delay_drop.dbg = dbg;
-
-	dbg->dir_debugfs =
-		debugfs_create_dir("delay_drop",
-				   dev->mdev->priv.dbg_root);
-	if (!dbg->dir_debugfs)
-		goto out_debugfs;
-
-	dbg->events_cnt_debugfs =
-		debugfs_create_atomic_t("num_timeout_events", 0400,
-					dbg->dir_debugfs,
-					&dev->delay_drop.events_cnt);
-	if (!dbg->events_cnt_debugfs)
-		goto out_debugfs;
-
-	dbg->rqs_cnt_debugfs =
-		debugfs_create_atomic_t("num_rqs", 0400,
-					dbg->dir_debugfs,
-					&dev->delay_drop.rqs_cnt);
-	if (!dbg->rqs_cnt_debugfs)
-		goto out_debugfs;
-
-	dbg->timeout_debugfs =
-		debugfs_create_file("timeout", 0600,
-				    dbg->dir_debugfs,
-				    &dev->delay_drop,
-				    &fops_delay_drop_timeout);
-	if (!dbg->timeout_debugfs)
-		goto out_debugfs;
+		return;
 
-	return 0;
+	root = debugfs_create_dir("delay_drop", dev->mdev->priv.dbg_root);
+	dev->delay_drop.dir_debugfs = root;
 
-out_debugfs:
-	delay_drop_debugfs_cleanup(dev);
-	return -ENOMEM;
+	debugfs_create_atomic_t("num_timeout_events", 0400, root,
+				&dev->delay_drop.events_cnt);
+	debugfs_create_atomic_t("num_rqs", 0400, root,
+				&dev->delay_drop.rqs_cnt);
+	debugfs_create_file("timeout", 0600, root, &dev->delay_drop,
+			    &fops_delay_drop_timeout);
 }
 
 static void init_delay_drop(struct mlx5_ib_dev *dev)
@@ -5826,8 +5786,7 @@ static void init_delay_drop(struct mlx5_ib_dev *dev)
 	atomic_set(&dev->delay_drop.rqs_cnt, 0);
 	atomic_set(&dev->delay_drop.events_cnt, 0);
 
-	if (delay_drop_debugfs_init(dev))
-		mlx5_ib_warn(dev, "Failed to init delay drop debugfs\n");
+	delay_drop_debugfs_init(dev);
 }
 
 static void mlx5_ib_unbind_slave_port(struct mlx5_ib_dev *ibdev,
@@ -6145,11 +6104,10 @@ static struct ib_counters *mlx5_ib_create_counters(struct ib_device *device,
 static void mlx5_ib_stage_init_cleanup(struct mlx5_ib_dev *dev)
 {
 	mlx5_ib_cleanup_multiport_master(dev);
-	if (IS_ENABLED(CONFIG_INFINIBAND_ON_DEMAND_PAGING)) {
-		srcu_barrier(&dev->mr_srcu);
-		cleanup_srcu_struct(&dev->mr_srcu);
-	}
+	WARN_ON(!xa_empty(&dev->odp_mkeys));
+	cleanup_srcu_struct(&dev->odp_srcu);
 
+	WARN_ON(!xa_empty(&dev->sig_mrs));
 	WARN_ON(!bitmap_empty(dev->dm.memic_alloc_pages, MLX5_MAX_MEMIC_PAGES));
 }
 
@@ -6201,15 +6159,15 @@ static int mlx5_ib_stage_init_init(struct mlx5_ib_dev *dev)
 	mutex_init(&dev->cap_mask_mutex);
 	INIT_LIST_HEAD(&dev->qp_list);
 	spin_lock_init(&dev->reset_flow_resource_lock);
+	xa_init(&dev->odp_mkeys);
+	xa_init(&dev->sig_mrs);
 
 	spin_lock_init(&dev->dm.lock);
 	dev->dm.dev = mdev;
 
-	if (IS_ENABLED(CONFIG_INFINIBAND_ON_DEMAND_PAGING)) {
-		err = init_srcu_struct(&dev->mr_srcu);
-		if (err)
-			goto err_mp;
-	}
+	err = init_srcu_struct(&dev->odp_srcu);
+	if (err)
+		goto err_mp;
 
 	return 0;
 
@@ -6269,6 +6227,9 @@ static const struct ib_device_ops mlx5_ib_dev_ops = {
 	.disassociate_ucontext = mlx5_ib_disassociate_ucontext,
 	.drain_rq = mlx5_ib_drain_rq,
 	.drain_sq = mlx5_ib_drain_sq,
+	.enable_driver = mlx5_ib_enable_driver,
+	.fill_res_entry = mlx5_ib_fill_res_entry,
+	.fill_stat_entry = mlx5_ib_fill_stat_entry,
 	.get_dev_fw_str = get_dev_fw_str,
 	.get_dma_mr = mlx5_ib_get_dma_mr,
 	.get_link_layer = mlx5_ib_port_link_layer,
@@ -6315,6 +6276,7 @@ static const struct ib_device_ops mlx5_ib_dev_ipoib_enhanced_ops = {
 
 static const struct ib_device_ops mlx5_ib_dev_sriov_ops = {
 	.get_vf_config = mlx5_ib_get_vf_config,
+	.get_vf_guid = mlx5_ib_get_vf_guid,
 	.get_vf_stats = mlx5_ib_get_vf_stats,
 	.set_vf_guid = mlx5_ib_set_vf_guid,
 	.set_vf_link_state = mlx5_ib_set_vf_link_state,
@@ -6444,7 +6406,7 @@ static const struct ib_device_ops mlx5_ib_dev_port_rep_ops = {
 	.query_port = mlx5_ib_rep_query_port,
 };
 
-static int mlx5_ib_stage_rep_non_default_cb(struct mlx5_ib_dev *dev)
+static int mlx5_ib_stage_raw_eth_non_default_cb(struct mlx5_ib_dev *dev)
 {
 	ib_set_device_ops(&dev->ib_dev, &mlx5_ib_dev_port_rep_ops);
 	return 0;
@@ -6484,7 +6446,7 @@ static void mlx5_ib_stage_common_roce_cleanup(struct mlx5_ib_dev *dev)
 	mlx5_remove_netdev_notifier(dev, port_num);
 }
 
-static int mlx5_ib_stage_rep_roce_init(struct mlx5_ib_dev *dev)
+static int mlx5_ib_stage_raw_eth_roce_init(struct mlx5_ib_dev *dev)
 {
 	struct mlx5_core_dev *mdev = dev->mdev;
 	enum rdma_link_layer ll;
@@ -6500,7 +6462,7 @@ static int mlx5_ib_stage_rep_roce_init(struct mlx5_ib_dev *dev)
 	return err;
 }
 
-static void mlx5_ib_stage_rep_roce_cleanup(struct mlx5_ib_dev *dev)
+static void mlx5_ib_stage_raw_eth_roce_cleanup(struct mlx5_ib_dev *dev)
 {
 	mlx5_ib_stage_common_roce_cleanup(dev);
 }
@@ -6710,6 +6672,18 @@ static void mlx5_ib_stage_devx_cleanup(struct mlx5_ib_dev *dev)
 	}
 }
 
+int mlx5_ib_enable_driver(struct ib_device *dev)
+{
+	struct mlx5_ib_dev *mdev = to_mdev(dev);
+	int ret;
+
+	ret = mlx5_ib_test_wc(mdev);
+	mlx5_ib_dbg(mdev, "Write-Combining %s",
+		    mdev->wc_support ? "supported" : "not supported");
+
+	return ret;
+}
+
 void __mlx5_ib_remove(struct mlx5_ib_dev *dev,
 		      const struct mlx5_ib_profile *profile,
 		      int stage)
@@ -6807,7 +6781,7 @@ static const struct mlx5_ib_profile pf_profile = {
 		     mlx5_ib_stage_delay_drop_cleanup),
 };
 
-const struct mlx5_ib_profile uplink_rep_profile = {
+const struct mlx5_ib_profile raw_eth_profile = {
 	STAGE_CREATE(MLX5_IB_STAGE_INIT,
 		     mlx5_ib_stage_init_init,
 		     mlx5_ib_stage_init_cleanup),
@@ -6818,11 +6792,11 @@ const struct mlx5_ib_profile uplink_rep_profile = {
 		     mlx5_ib_stage_caps_init,
 		     NULL),
 	STAGE_CREATE(MLX5_IB_STAGE_NON_DEFAULT_CB,
-		     mlx5_ib_stage_rep_non_default_cb,
+		     mlx5_ib_stage_raw_eth_non_default_cb,
 		     NULL),
 	STAGE_CREATE(MLX5_IB_STAGE_ROCE,
-		     mlx5_ib_stage_rep_roce_init,
-		     mlx5_ib_stage_rep_roce_cleanup),
+		     mlx5_ib_stage_raw_eth_roce_init,
+		     mlx5_ib_stage_raw_eth_roce_cleanup),
 	STAGE_CREATE(MLX5_IB_STAGE_SRQ,
 		     mlx5_init_srq_table,
 		     mlx5_cleanup_srq_table),
@@ -6898,6 +6872,7 @@ static void *mlx5_ib_add_slave_port(struct mlx5_core_dev *mdev)
 
 static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 {
+	const struct mlx5_ib_profile *profile;
 	enum rdma_link_layer ll;
 	struct mlx5_ib_dev *dev;
 	int port_type_cap;
@@ -6933,7 +6908,12 @@ static void *mlx5_ib_add(struct mlx5_core_dev *mdev)
 	dev->mdev = mdev;
 	dev->num_ports = num_ports;
 
-	return __mlx5_ib_add(dev, &pf_profile);
+	if (ll == IB_LINK_LAYER_ETHERNET && !mlx5_is_roce_enabled(mdev))
+		profile = &raw_eth_profile;
+	else
+		profile = &pf_profile;
+
+	return __mlx5_ib_add(dev, profile);
 }
 
 static void mlx5_ib_remove(struct mlx5_core_dev *mdev, void *context)