net/mlx5: Change the acl enable prototype to return status
authorMohamad Haj Yahia <mohamad@mellanox.com>
Tue, 25 Oct 2016 15:36:31 +0000 (18:36 +0300)
committerDavid S. Miller <davem@davemloft.net>
Sat, 29 Oct 2016 16:00:39 +0000 (12:00 -0400)
The Ingress/Egress ACL enable function may fail and it should return
status to its caller to avoid NULL pointer dereference.

Fixes: f942380c1239 ('net/mlx5: E-Switch, Vport ingress/egress ACLs rules for spoofchk')
Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/mellanox/mlx5/core/eswitch.c

index abbf2c369923d02534b1ebfa82e212bcefce0333..be1f7333ab7f6845acc01ae46fd9ba150f6ad6aa 100644 (file)
@@ -931,8 +931,8 @@ static void esw_vport_change_handler(struct work_struct *work)
        mutex_unlock(&esw->state_lock);
 }
 
-static void esw_vport_enable_egress_acl(struct mlx5_eswitch *esw,
-                                       struct mlx5_vport *vport)
+static int esw_vport_enable_egress_acl(struct mlx5_eswitch *esw,
+                                      struct mlx5_vport *vport)
 {
        int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
        struct mlx5_flow_group *vlan_grp = NULL;
@@ -949,9 +949,11 @@ static void esw_vport_enable_egress_acl(struct mlx5_eswitch *esw,
        int table_size = 2;
        int err = 0;
 
-       if (!MLX5_CAP_ESW_EGRESS_ACL(dev, ft_support) ||
-           !IS_ERR_OR_NULL(vport->egress.acl))
-               return;
+       if (!MLX5_CAP_ESW_EGRESS_ACL(dev, ft_support))
+               return -EOPNOTSUPP;
+
+       if (!IS_ERR_OR_NULL(vport->egress.acl))
+               return 0;
 
        esw_debug(dev, "Create vport[%d] egress ACL log_max_size(%d)\n",
                  vport->vport, MLX5_CAP_ESW_EGRESS_ACL(dev, log_max_ft_size));
@@ -959,12 +961,12 @@ static void esw_vport_enable_egress_acl(struct mlx5_eswitch *esw,
        root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_ESW_EGRESS);
        if (!root_ns) {
                esw_warn(dev, "Failed to get E-Switch egress flow namespace\n");
-               return;
+               return -EIO;
        }
 
        flow_group_in = mlx5_vzalloc(inlen);
        if (!flow_group_in)
-               return;
+               return -ENOMEM;
 
        acl = mlx5_create_vport_flow_table(root_ns, 0, table_size, 0, vport->vport);
        if (IS_ERR(acl)) {
@@ -1009,6 +1011,7 @@ out:
                mlx5_destroy_flow_group(vlan_grp);
        if (err && !IS_ERR_OR_NULL(acl))
                mlx5_destroy_flow_table(acl);
+       return err;
 }
 
 static void esw_vport_cleanup_egress_rules(struct mlx5_eswitch *esw,
@@ -1041,8 +1044,8 @@ static void esw_vport_disable_egress_acl(struct mlx5_eswitch *esw,
        vport->egress.acl = NULL;
 }
 
-static void esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw,
-                                        struct mlx5_vport *vport)
+static int esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw,
+                                       struct mlx5_vport *vport)
 {
        int inlen = MLX5_ST_SZ_BYTES(create_flow_group_in);
        struct mlx5_core_dev *dev = esw->dev;
@@ -1063,9 +1066,11 @@ static void esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw,
        int table_size = 4;
        int err = 0;
 
-       if (!MLX5_CAP_ESW_INGRESS_ACL(dev, ft_support) ||
-           !IS_ERR_OR_NULL(vport->ingress.acl))
-               return;
+       if (!MLX5_CAP_ESW_INGRESS_ACL(dev, ft_support))
+               return -EOPNOTSUPP;
+
+       if (!IS_ERR_OR_NULL(vport->ingress.acl))
+               return 0;
 
        esw_debug(dev, "Create vport[%d] ingress ACL log_max_size(%d)\n",
                  vport->vport, MLX5_CAP_ESW_INGRESS_ACL(dev, log_max_ft_size));
@@ -1073,12 +1078,12 @@ static void esw_vport_enable_ingress_acl(struct mlx5_eswitch *esw,
        root_ns = mlx5_get_flow_namespace(dev, MLX5_FLOW_NAMESPACE_ESW_INGRESS);
        if (!root_ns) {
                esw_warn(dev, "Failed to get E-Switch ingress flow namespace\n");
-               return;
+               return -EIO;
        }
 
        flow_group_in = mlx5_vzalloc(inlen);
        if (!flow_group_in)
-               return;
+               return -ENOMEM;
 
        acl = mlx5_create_vport_flow_table(root_ns, 0, table_size, 0, vport->vport);
        if (IS_ERR(acl)) {
@@ -1167,6 +1172,7 @@ out:
        }
 
        kvfree(flow_group_in);
+       return err;
 }
 
 static void esw_vport_cleanup_ingress_rules(struct mlx5_eswitch *esw,
@@ -1225,7 +1231,13 @@ static int esw_vport_ingress_config(struct mlx5_eswitch *esw,
                return 0;
        }
 
-       esw_vport_enable_ingress_acl(esw, vport);
+       err = esw_vport_enable_ingress_acl(esw, vport);
+       if (err) {
+               mlx5_core_warn(esw->dev,
+                              "failed to enable ingress acl (%d) on vport[%d]\n",
+                              err, vport->vport);
+               return err;
+       }
 
        esw_debug(esw->dev,
                  "vport[%d] configure ingress rules, vlan(%d) qos(%d)\n",
@@ -1299,7 +1311,13 @@ static int esw_vport_egress_config(struct mlx5_eswitch *esw,
                return 0;
        }
 
-       esw_vport_enable_egress_acl(esw, vport);
+       err = esw_vport_enable_egress_acl(esw, vport);
+       if (err) {
+               mlx5_core_warn(esw->dev,
+                              "failed to enable egress acl (%d) on vport[%d]\n",
+                              err, vport->vport);
+               return err;
+       }
 
        esw_debug(esw->dev,
                  "vport[%d] configure egress rules, vlan(%d) qos(%d)\n",